ods a3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "from pathlib2 import Path\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from scipy.sparse import csr_matrix, hstack\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from matplotlib import pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set()\n",
    "%config InlineBackend.figure_format = 'retina'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from numpy import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>...</th>\n",
       "      <th>time6</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>251175</th>\n",
       "      <td>270</td>\n",
       "      <td>2013-02-12 16:25:10</td>\n",
       "      <td>270.0</td>\n",
       "      <td>2013-02-12 16:25:11</td>\n",
       "      <td>270.0</td>\n",
       "      <td>2013-02-12 16:32:10</td>\n",
       "      <td>21.0</td>\n",
       "      <td>2013-02-12 16:32:11</td>\n",
       "      <td>21.0</td>\n",
       "      <td>2013-02-12 16:32:24</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-02-12 16:32:25</td>\n",
       "      <td>21.0</td>\n",
       "      <td>2013-02-12 16:32:25</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:32:26</td>\n",
       "      <td>30.0</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196388</th>\n",
       "      <td>29</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:32:28</td>\n",
       "      <td>37.0</td>\n",
       "      <td>2013-02-12 16:32:29</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:32:34</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:32:35</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-02-12 16:32:35</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:32:42</td>\n",
       "      <td>29.0</td>\n",
       "      <td>2013-02-12 16:32:42</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:32:51</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172448</th>\n",
       "      <td>29</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:33:11</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:33:12</td>\n",
       "      <td>29.0</td>\n",
       "      <td>2013-02-12 16:33:13</td>\n",
       "      <td>37.0</td>\n",
       "      <td>2013-02-12 16:33:15</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-02-12 16:33:24</td>\n",
       "      <td>29.0</td>\n",
       "      <td>2013-02-12 16:33:24</td>\n",
       "      <td>7832.0</td>\n",
       "      <td>2013-02-12 16:33:33</td>\n",
       "      <td>29.0</td>\n",
       "      <td>2013-02-12 16:33:34</td>\n",
       "      <td>270.0</td>\n",
       "      <td>2013-02-12 16:33:46</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70129</th>\n",
       "      <td>167</td>\n",
       "      <td>2013-02-12 16:33:50</td>\n",
       "      <td>167.0</td>\n",
       "      <td>2013-02-12 16:33:51</td>\n",
       "      <td>1515.0</td>\n",
       "      <td>2013-02-12 16:33:52</td>\n",
       "      <td>167.0</td>\n",
       "      <td>2013-02-12 16:33:52</td>\n",
       "      <td>37.0</td>\n",
       "      <td>2013-02-12 16:33:52</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-02-12 16:33:52</td>\n",
       "      <td>855.0</td>\n",
       "      <td>2013-02-12 16:33:52</td>\n",
       "      <td>1515.0</td>\n",
       "      <td>2013-02-12 16:33:53</td>\n",
       "      <td>855.0</td>\n",
       "      <td>2013-02-12 16:33:53</td>\n",
       "      <td>1514.0</td>\n",
       "      <td>2013-02-12 16:33:53</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206254</th>\n",
       "      <td>1520</td>\n",
       "      <td>2013-02-12 16:33:55</td>\n",
       "      <td>1522.0</td>\n",
       "      <td>2013-02-12 16:33:56</td>\n",
       "      <td>1522.0</td>\n",
       "      <td>2013-02-12 16:34:01</td>\n",
       "      <td>1515.0</td>\n",
       "      <td>2013-02-12 16:34:12</td>\n",
       "      <td>1515.0</td>\n",
       "      <td>2013-02-12 16:34:13</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-02-12 16:34:24</td>\n",
       "      <td>1514.0</td>\n",
       "      <td>2013-02-12 16:34:24</td>\n",
       "      <td>1515.0</td>\n",
       "      <td>2013-02-12 16:34:24</td>\n",
       "      <td>1520.0</td>\n",
       "      <td>2013-02-12 16:34:25</td>\n",
       "      <td>1521.0</td>\n",
       "      <td>2013-02-12 16:34:25</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1   site2               time2   site3  \\\n",
       "session_id                                                                  \n",
       "251175        270 2013-02-12 16:25:10   270.0 2013-02-12 16:25:11   270.0   \n",
       "196388         29 2013-02-12 16:32:27  7832.0 2013-02-12 16:32:28    37.0   \n",
       "172448         29 2013-02-12 16:32:53  7832.0 2013-02-12 16:33:11  7832.0   \n",
       "70129         167 2013-02-12 16:33:50   167.0 2013-02-12 16:33:51  1515.0   \n",
       "206254       1520 2013-02-12 16:33:55  1522.0 2013-02-12 16:33:56  1522.0   \n",
       "\n",
       "                         time3   site4               time4   site5  \\\n",
       "session_id                                                           \n",
       "251175     2013-02-12 16:32:10    21.0 2013-02-12 16:32:11    21.0   \n",
       "196388     2013-02-12 16:32:29  7832.0 2013-02-12 16:32:34  7832.0   \n",
       "172448     2013-02-12 16:33:12    29.0 2013-02-12 16:33:13    37.0   \n",
       "70129      2013-02-12 16:33:52   167.0 2013-02-12 16:33:52    37.0   \n",
       "206254     2013-02-12 16:34:01  1515.0 2013-02-12 16:34:12  1515.0   \n",
       "\n",
       "                         time5  ...               time6   site7  \\\n",
       "session_id                      ...                               \n",
       "251175     2013-02-12 16:32:24  ... 2013-02-12 16:32:25    21.0   \n",
       "196388     2013-02-12 16:32:35  ... 2013-02-12 16:32:35  7832.0   \n",
       "172448     2013-02-12 16:33:15  ... 2013-02-12 16:33:24    29.0   \n",
       "70129      2013-02-12 16:33:52  ... 2013-02-12 16:33:52   855.0   \n",
       "206254     2013-02-12 16:34:13  ... 2013-02-12 16:34:24  1514.0   \n",
       "\n",
       "                         time7   site8               time8   site9  \\\n",
       "session_id                                                           \n",
       "251175     2013-02-12 16:32:25  7832.0 2013-02-12 16:32:26    30.0   \n",
       "196388     2013-02-12 16:32:42    29.0 2013-02-12 16:32:42  7832.0   \n",
       "172448     2013-02-12 16:33:24  7832.0 2013-02-12 16:33:33    29.0   \n",
       "70129      2013-02-12 16:33:52  1515.0 2013-02-12 16:33:53   855.0   \n",
       "206254     2013-02-12 16:34:24  1515.0 2013-02-12 16:34:24  1520.0   \n",
       "\n",
       "                         time9  site10              time10 target  \n",
       "session_id                                                         \n",
       "251175     2013-02-12 16:32:27  7832.0 2013-02-12 16:32:27      1  \n",
       "196388     2013-02-12 16:32:51  7832.0 2013-02-12 16:32:53      1  \n",
       "172448     2013-02-12 16:33:34   270.0 2013-02-12 16:33:46      1  \n",
       "70129      2013-02-12 16:33:53  1514.0 2013-02-12 16:33:53      1  \n",
       "206254     2013-02-12 16:34:25  1521.0 2013-02-12 16:34:25      1  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "PATH_TO_DATA = Path('catch-me-if-you-can-intruder-detection-through-webpage-session-tracking2/')\n",
    "\n",
    "times = ['time%s' % i for i in range(1, 11)]\n",
    "train_df = pd.read_csv(PATH_TO_DATA / 'train_sessions.csv',\n",
    "                       index_col='session_id', parse_dates=times)\n",
    "test_df = pd.read_csv(PATH_TO_DATA / 'test_sessions.csv',\n",
    "                      index_col='session_id', parse_dates=times)\n",
    "\n",
    "# Sort the data by time\n",
    "train_df = train_df.sort_values(by='time1')\n",
    "\n",
    "# Look at the first rows of the training set\n",
    "train_alice = train_df[train_df['target']== 1]\n",
    "#test_alice = test_df[test_df['target']== 1]\n",
    "train_alice.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((253561, 21), (82797, 20))"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.shape, test_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['site1', 'time1', 'site2', 'time2', 'site3', 'time3', 'site4', 'time4',\n",
       "       'site5', 'time5', 'site6', 'time6', 'site7', 'time7', 'site8', 'time8',\n",
       "       'site9', 'time9', 'site10', 'time10'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['site1', 'time1', 'site2', 'time2', 'site3', 'time3', 'site4', 'time4',\n",
       "       'site5', 'time5', 'site6', 'time6', 'site7', 'time7', 'site8', 'time8',\n",
       "       'site9', 'time9', 'site10', 'time10', 'target'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    251264\n",
       "1      2297\n",
       "Name: target, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.target.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>site6</th>\n",
       "      <th>time6</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>29</td>\n",
       "      <td>2014-10-04 11:19:53</td>\n",
       "      <td>35.0</td>\n",
       "      <td>2014-10-04 11:19:53</td>\n",
       "      <td>22.0</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>321.0</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>23.0</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>2211.0</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>6730.0</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>21.0</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>44582.0</td>\n",
       "      <td>2014-10-04 11:20:00</td>\n",
       "      <td>15336.0</td>\n",
       "      <td>2014-10-04 11:20:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:00:28</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:00:53</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:00:58</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:01:06</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:01:09</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:01:10</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:01:23</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:01:29</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:01:30</td>\n",
       "      <td>782.0</td>\n",
       "      <td>2014-07-03 11:01:53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>55</td>\n",
       "      <td>2014-12-05 15:55:12</td>\n",
       "      <td>55.0</td>\n",
       "      <td>2014-12-05 15:55:13</td>\n",
       "      <td>55.0</td>\n",
       "      <td>2014-12-05 15:55:14</td>\n",
       "      <td>55.0</td>\n",
       "      <td>2014-12-05 15:56:15</td>\n",
       "      <td>55.0</td>\n",
       "      <td>2014-12-05 15:56:16</td>\n",
       "      <td>55.0</td>\n",
       "      <td>2014-12-05 15:56:17</td>\n",
       "      <td>55.0</td>\n",
       "      <td>2014-12-05 15:56:18</td>\n",
       "      <td>55.0</td>\n",
       "      <td>2014-12-05 15:56:19</td>\n",
       "      <td>1445.0</td>\n",
       "      <td>2014-12-05 15:56:33</td>\n",
       "      <td>1445.0</td>\n",
       "      <td>2014-12-05 15:56:36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1023</td>\n",
       "      <td>2014-11-04 10:03:19</td>\n",
       "      <td>1022.0</td>\n",
       "      <td>2014-11-04 10:03:19</td>\n",
       "      <td>50.0</td>\n",
       "      <td>2014-11-04 10:03:20</td>\n",
       "      <td>222.0</td>\n",
       "      <td>2014-11-04 10:03:21</td>\n",
       "      <td>202.0</td>\n",
       "      <td>2014-11-04 10:03:21</td>\n",
       "      <td>3374.0</td>\n",
       "      <td>2014-11-04 10:03:22</td>\n",
       "      <td>50.0</td>\n",
       "      <td>2014-11-04 10:03:22</td>\n",
       "      <td>48.0</td>\n",
       "      <td>2014-11-04 10:03:22</td>\n",
       "      <td>48.0</td>\n",
       "      <td>2014-11-04 10:03:23</td>\n",
       "      <td>3374.0</td>\n",
       "      <td>2014-11-04 10:03:23</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1   site2               time2  site3  \\\n",
       "session_id                                                                 \n",
       "1              29 2014-10-04 11:19:53    35.0 2014-10-04 11:19:53   22.0   \n",
       "2             782 2014-07-03 11:00:28   782.0 2014-07-03 11:00:53  782.0   \n",
       "3              55 2014-12-05 15:55:12    55.0 2014-12-05 15:55:13   55.0   \n",
       "4            1023 2014-11-04 10:03:19  1022.0 2014-11-04 10:03:19   50.0   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "1          2014-10-04 11:19:54  321.0 2014-10-04 11:19:54   23.0   \n",
       "2          2014-07-03 11:00:58  782.0 2014-07-03 11:01:06  782.0   \n",
       "3          2014-12-05 15:55:14   55.0 2014-12-05 15:56:15   55.0   \n",
       "4          2014-11-04 10:03:20  222.0 2014-11-04 10:03:21  202.0   \n",
       "\n",
       "                         time5   site6               time6   site7  \\\n",
       "session_id                                                           \n",
       "1          2014-10-04 11:19:54  2211.0 2014-10-04 11:19:54  6730.0   \n",
       "2          2014-07-03 11:01:09   782.0 2014-07-03 11:01:10   782.0   \n",
       "3          2014-12-05 15:56:16    55.0 2014-12-05 15:56:17    55.0   \n",
       "4          2014-11-04 10:03:21  3374.0 2014-11-04 10:03:22    50.0   \n",
       "\n",
       "                         time7  site8               time8    site9  \\\n",
       "session_id                                                           \n",
       "1          2014-10-04 11:19:54   21.0 2014-10-04 11:19:54  44582.0   \n",
       "2          2014-07-03 11:01:23  782.0 2014-07-03 11:01:29    782.0   \n",
       "3          2014-12-05 15:56:18   55.0 2014-12-05 15:56:19   1445.0   \n",
       "4          2014-11-04 10:03:22   48.0 2014-11-04 10:03:22     48.0   \n",
       "\n",
       "                         time9   site10              time10  \n",
       "session_id                                                   \n",
       "1          2014-10-04 11:20:00  15336.0 2014-10-04 11:20:00  \n",
       "2          2014-07-03 11:01:30    782.0 2014-07-03 11:01:53  \n",
       "3          2014-12-05 15:56:33   1445.0 2014-12-05 15:56:36  \n",
       "4          2014-11-04 10:03:23   3374.0 2014-11-04 10:03:23  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df.head(4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>site2</th>\n",
       "      <th>site3</th>\n",
       "      <th>site4</th>\n",
       "      <th>site5</th>\n",
       "      <th>site6</th>\n",
       "      <th>site7</th>\n",
       "      <th>site8</th>\n",
       "      <th>site9</th>\n",
       "      <th>site10</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>2297.000000</td>\n",
       "      <td>2294.000000</td>\n",
       "      <td>2287.000000</td>\n",
       "      <td>2286.000000</td>\n",
       "      <td>2280.000000</td>\n",
       "      <td>2273.000000</td>\n",
       "      <td>2269.000000</td>\n",
       "      <td>2263.000000</td>\n",
       "      <td>2262.000000</td>\n",
       "      <td>2258.000000</td>\n",
       "      <td>2297.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>3119.592947</td>\n",
       "      <td>2999.513078</td>\n",
       "      <td>2951.274158</td>\n",
       "      <td>3072.664042</td>\n",
       "      <td>2912.150439</td>\n",
       "      <td>3035.263968</td>\n",
       "      <td>3176.235346</td>\n",
       "      <td>2930.432170</td>\n",
       "      <td>3089.833333</td>\n",
       "      <td>2981.442870</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>7120.997108</td>\n",
       "      <td>6644.107233</td>\n",
       "      <td>6628.797591</td>\n",
       "      <td>6825.233710</td>\n",
       "      <td>6648.011922</td>\n",
       "      <td>6765.622736</td>\n",
       "      <td>6953.305236</td>\n",
       "      <td>6723.534647</td>\n",
       "      <td>6873.117594</td>\n",
       "      <td>6660.374884</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>76.000000</td>\n",
       "      <td>76.000000</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>76.000000</td>\n",
       "      <td>75.750000</td>\n",
       "      <td>76.000000</td>\n",
       "      <td>76.000000</td>\n",
       "      <td>76.000000</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>76.000000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>170.000000</td>\n",
       "      <td>252.000000</td>\n",
       "      <td>162.000000</td>\n",
       "      <td>261.000000</td>\n",
       "      <td>143.500000</td>\n",
       "      <td>177.000000</td>\n",
       "      <td>240.000000</td>\n",
       "      <td>143.000000</td>\n",
       "      <td>144.500000</td>\n",
       "      <td>162.000000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1519.000000</td>\n",
       "      <td>1677.000000</td>\n",
       "      <td>1522.000000</td>\n",
       "      <td>1797.500000</td>\n",
       "      <td>1521.250000</td>\n",
       "      <td>1523.000000</td>\n",
       "      <td>2077.000000</td>\n",
       "      <td>1519.000000</td>\n",
       "      <td>1569.000000</td>\n",
       "      <td>1520.000000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>27381.000000</td>\n",
       "      <td>27387.000000</td>\n",
       "      <td>27381.000000</td>\n",
       "      <td>27379.000000</td>\n",
       "      <td>27389.000000</td>\n",
       "      <td>27389.000000</td>\n",
       "      <td>27389.000000</td>\n",
       "      <td>27388.000000</td>\n",
       "      <td>27389.000000</td>\n",
       "      <td>27381.000000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              site1         site2         site3         site4         site5  \\\n",
       "count   2297.000000   2294.000000   2287.000000   2286.000000   2280.000000   \n",
       "mean    3119.592947   2999.513078   2951.274158   3072.664042   2912.150439   \n",
       "std     7120.997108   6644.107233   6628.797591   6825.233710   6648.011922   \n",
       "min        1.000000      3.000000      1.000000      2.000000      2.000000   \n",
       "25%       76.000000     76.000000     75.000000     76.000000     75.750000   \n",
       "50%      170.000000    252.000000    162.000000    261.000000    143.500000   \n",
       "75%     1519.000000   1677.000000   1522.000000   1797.500000   1521.250000   \n",
       "max    27381.000000  27387.000000  27381.000000  27379.000000  27389.000000   \n",
       "\n",
       "              site6         site7         site8         site9        site10  \\\n",
       "count   2273.000000   2269.000000   2263.000000   2262.000000   2258.000000   \n",
       "mean    3035.263968   3176.235346   2930.432170   3089.833333   2981.442870   \n",
       "std     6765.622736   6953.305236   6723.534647   6873.117594   6660.374884   \n",
       "min        1.000000      2.000000      1.000000      1.000000      3.000000   \n",
       "25%       76.000000     76.000000     76.000000     75.000000     76.000000   \n",
       "50%      177.000000    240.000000    143.000000    144.500000    162.000000   \n",
       "75%     1523.000000   2077.000000   1519.000000   1569.000000   1520.000000   \n",
       "max    27389.000000  27389.000000  27388.000000  27389.000000  27381.000000   \n",
       "\n",
       "       target  \n",
       "count  2297.0  \n",
       "mean      1.0  \n",
       "std       0.0  \n",
       "min       1.0  \n",
       "25%       1.0  \n",
       "50%       1.0  \n",
       "75%       1.0  \n",
       "max       1.0  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_alice.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\pandas\\core\\frame.py:3391: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  self[k1] = value[k2]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Websites total: 48371\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>25075</th>\n",
       "      <td>www.abmecatronique.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13997</th>\n",
       "      <td>groups.live.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42436</th>\n",
       "      <td>majeureliguefootball.wordpress.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30911</th>\n",
       "      <td>cdt46.media.tourinsoft.eu</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8104</th>\n",
       "      <td>www.hdwallpapers.eu</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     site\n",
       "25075              www.abmecatronique.com\n",
       "13997                     groups.live.com\n",
       "42436  majeureliguefootball.wordpress.com\n",
       "30911           cdt46.media.tourinsoft.eu\n",
       "8104                  www.hdwallpapers.eu"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Change site1, ..., site10 columns type to integer and fill NA-values with zeros\n",
    "sites = ['site%s' % i for i in range(1, 11)]\n",
    "train_df[sites] = train_df[sites].fillna(0).astype(np.uint16)\n",
    "train_alice[sites] = train_alice[sites].fillna(0).astype(np.uint16)\n",
    "test_df[sites] = test_df[sites].fillna(0).astype(np.uint16)\n",
    "\n",
    "# Load websites dictionary\n",
    "with open(PATH_TO_DATA / 'site_dic.pkl', \"rb\") as input_file:\n",
    "    site_dict = pickle.load(input_file)\n",
    "\n",
    "# Create dataframe for the dictionary\n",
    "sites_dict = pd.DataFrame(list(site_dict.keys()), index=list(site_dict.values()), \n",
    "                          columns=['site'])\n",
    "print(u'Websites total:', sites_dict.shape[0])\n",
    "sites_dict.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "site    i1.ytimg.com\n",
       "Name: 77, dtype: object"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sites_dict.loc[77]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(48371, 2297)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(sites_dict), len(train_alice)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "77    1382\n",
      "80    1354\n",
      "76    1307\n",
      "29     897\n",
      "21     857\n",
      "dtype: int64\n"
     ]
    }
   ],
   "source": [
    "top_five_sites_for_alice = pd.Series(train_alice[sites].values.flatten()\n",
    "                     ).value_counts().sort_values(ascending=False).head(5)\n",
    "print(top_five_sites_for_alice)\n",
    "#sites_dict.loc[top_five_sites_for_alice.drop(0).index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.series.Series'>\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(1054, None)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_sites = pd.Series(train_alice[sites].values.flatten()).value_counts()\n",
    "len(top_sites), print(type(top_sites))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "top_sites= pd.DataFrame(top_sites)\n",
    "top_sites.rename({0:'num_visits'}, inplace=True, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>num_visits</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>77</th>\n",
       "      <td>1382</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80</th>\n",
       "      <td>1354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>1307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>857</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    num_visits\n",
       "77        1382\n",
       "80        1354\n",
       "76        1307\n",
       "29         897\n",
       "21         857"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_sites.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "top_sites = top_sites.sort_index(ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>num_visits</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   num_visits\n",
       "6          35\n",
       "4          26\n",
       "3         129\n",
       "2          18\n",
       "1          12"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_sites.drop([0],axis=0, inplace=True)\n",
    "top_sites.tail(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>num_visits</th>\n",
       "      <th>site</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>26</td>\n",
       "      <td>storage.live.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>48</td>\n",
       "      <td>go.trouter.io</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>14</td>\n",
       "      <td>js.live.net</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>80</td>\n",
       "      <td>windowslive.tt.omtrdc.net</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>26</td>\n",
       "      <td>secure.shared.live.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>35</td>\n",
       "      <td>people.directory.live.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>26</td>\n",
       "      <td>mail.live.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>129</td>\n",
       "      <td>login.live.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>18</td>\n",
       "      <td>hotmail.fr</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>12</td>\n",
       "      <td>fpdownload2.macromedia.com</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    num_visits                        site\n",
       "12          26            storage.live.com\n",
       "11          48               go.trouter.io\n",
       "10          14                 js.live.net\n",
       "8           80   windowslive.tt.omtrdc.net\n",
       "7           26      secure.shared.live.com\n",
       "6           35   people.directory.live.com\n",
       "4           26               mail.live.com\n",
       "3          129              login.live.com\n",
       "2           18                  hotmail.fr\n",
       "1           12  fpdownload2.macromedia.com"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df3 = top_sites.join(sites_dict)\n",
    "df3.tail(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "df3.sort_values(by='num_visits',ascending=True,inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1053\n"
     ]
    }
   ],
   "source": [
    "print(len(df3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>num_visits</th>\n",
       "      <th>site</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>82</th>\n",
       "      <td>447</td>\n",
       "      <td>r2---sn-gxo5uxg-jqbe.googlevideo.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>451</td>\n",
       "      <td>s.ytimg.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>522</td>\n",
       "      <td>apis.google.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>879</th>\n",
       "      <td>522</td>\n",
       "      <td>r1---sn-gxo5uxg-jqbe.googlevideo.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81</th>\n",
       "      <td>609</td>\n",
       "      <td>r4---sn-gxo5uxg-jqbe.googlevideo.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>857</td>\n",
       "      <td>www.google.fr</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>897</td>\n",
       "      <td>www.facebook.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>1307</td>\n",
       "      <td>www.youtube.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80</th>\n",
       "      <td>1354</td>\n",
       "      <td>s.youtube.com</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77</th>\n",
       "      <td>1382</td>\n",
       "      <td>i1.ytimg.com</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     num_visits                                  site\n",
       "82          447  r2---sn-gxo5uxg-jqbe.googlevideo.com\n",
       "75          451                           s.ytimg.com\n",
       "22          522                       apis.google.com\n",
       "879         522  r1---sn-gxo5uxg-jqbe.googlevideo.com\n",
       "81          609  r4---sn-gxo5uxg-jqbe.googlevideo.com\n",
       "21          857                         www.google.fr\n",
       "29          897                      www.facebook.com\n",
       "76         1307                       www.youtube.com\n",
       "80         1354                         s.youtube.com\n",
       "77         1382                          i1.ytimg.com"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df3.tail(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>seconds</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>1786.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            target                 min                 max  seconds\n",
       "session_id                                                         \n",
       "21669            0 2013-01-12 08:05:57 2013-01-12 08:05:57      0.0\n",
       "54843            0 2013-01-12 08:37:23 2013-01-12 09:07:09   1786.0\n",
       "77292            0 2013-01-12 08:50:13 2013-01-12 08:50:17      4.0\n",
       "114021           0 2013-01-12 08:50:17 2013-01-12 08:50:20      3.0\n",
       "146670           0 2013-01-12 08:50:20 2013-01-12 08:50:22      2.0"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a separate dataframe where we will work with timestamps\n",
    "time_df = pd.DataFrame(index=train_df.index)\n",
    "time_df['target'] = train_df['target']\n",
    "\n",
    "# Find sessions' starting and ending\n",
    "time_df['min'] = train_df[times].min(axis=1)\n",
    "time_df['max'] = train_df[times].max(axis=1)\n",
    "\n",
    "# Calculate sessions' duration in seconds\n",
    "time_df['seconds'] = (time_df['max'] - time_df['min']) / np.timedelta64(1, 's')\n",
    "\n",
    "time_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>...</th>\n",
       "      <th>time6</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 09:07:07</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:14</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>784</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>242171</th>\n",
       "      <td>952</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:23</td>\n",
       "      <td>953</td>\n",
       "      <td>2013-01-12 08:50:23</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:23</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:24</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:24</td>\n",
       "      <td>953</td>\n",
       "      <td>2013-01-12 08:50:24</td>\n",
       "      <td>955</td>\n",
       "      <td>2013-01-12 08:50:24</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:25</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:25</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57157</th>\n",
       "      <td>953</td>\n",
       "      <td>2013-01-12 08:50:25</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:26</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:26</td>\n",
       "      <td>953</td>\n",
       "      <td>2013-01-12 08:50:26</td>\n",
       "      <td>955</td>\n",
       "      <td>2013-01-12 08:50:26</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:27</td>\n",
       "      <td>953</td>\n",
       "      <td>2013-01-12 08:50:27</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:27</td>\n",
       "      <td>953</td>\n",
       "      <td>2013-01-12 08:50:28</td>\n",
       "      <td>1033</td>\n",
       "      <td>2013-01-12 08:50:28</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>240201</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:28</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:28</td>\n",
       "      <td>954</td>\n",
       "      <td>2013-01-12 08:50:28</td>\n",
       "      <td>953</td>\n",
       "      <td>2013-01-12 08:50:29</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:29</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:29</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:30</td>\n",
       "      <td>956</td>\n",
       "      <td>2013-01-12 08:50:30</td>\n",
       "      <td>957</td>\n",
       "      <td>2013-01-12 08:50:31</td>\n",
       "      <td>956</td>\n",
       "      <td>2013-01-12 08:50:31</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>210686</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:31</td>\n",
       "      <td>956</td>\n",
       "      <td>2013-01-12 08:50:32</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:32</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:33</td>\n",
       "      <td>955</td>\n",
       "      <td>2013-01-12 08:50:33</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:33</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:34</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:35</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:36</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:36</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98804</th>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:37</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:37</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:38</td>\n",
       "      <td>784</td>\n",
       "      <td>2013-01-12 08:50:49</td>\n",
       "      <td>49</td>\n",
       "      <td>2013-01-12 08:50:59</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:51:03</td>\n",
       "      <td>812</td>\n",
       "      <td>2013-01-12 08:51:03</td>\n",
       "      <td>982</td>\n",
       "      <td>2013-01-12 08:51:03</td>\n",
       "      <td>52</td>\n",
       "      <td>2013-01-12 08:51:03</td>\n",
       "      <td>52</td>\n",
       "      <td>2013-01-12 08:51:04</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "21669          56 2013-01-12 08:05:57     55 2013-01-12 08:05:57      0   \n",
       "54843          56 2013-01-12 08:37:23     55 2013-01-12 08:37:23     56   \n",
       "77292         946 2013-01-12 08:50:13    946 2013-01-12 08:50:14    951   \n",
       "114021        945 2013-01-12 08:50:17    948 2013-01-12 08:50:17    949   \n",
       "146670        947 2013-01-12 08:50:20    950 2013-01-12 08:50:20    948   \n",
       "242171        952 2013-01-12 08:50:22    947 2013-01-12 08:50:23    953   \n",
       "57157         953 2013-01-12 08:50:25    947 2013-01-12 08:50:26    946   \n",
       "240201        946 2013-01-12 08:50:28    947 2013-01-12 08:50:28    954   \n",
       "210686        946 2013-01-12 08:50:31    956 2013-01-12 08:50:32    946   \n",
       "98804         948 2013-01-12 08:50:37    946 2013-01-12 08:50:37    948   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843      2013-01-12 09:07:07     55 2013-01-12 09:07:09      0   \n",
       "77292      2013-01-12 08:50:15    946 2013-01-12 08:50:15    946   \n",
       "114021     2013-01-12 08:50:18    948 2013-01-12 08:50:18    945   \n",
       "146670     2013-01-12 08:50:20    947 2013-01-12 08:50:21    950   \n",
       "242171     2013-01-12 08:50:23    946 2013-01-12 08:50:23    947   \n",
       "57157      2013-01-12 08:50:26    953 2013-01-12 08:50:26    955   \n",
       "240201     2013-01-12 08:50:28    953 2013-01-12 08:50:29    946   \n",
       "210686     2013-01-12 08:50:32    946 2013-01-12 08:50:33    955   \n",
       "98804      2013-01-12 08:50:38    784 2013-01-12 08:50:49     49   \n",
       "\n",
       "                         time5  ...               time6 site7  \\\n",
       "session_id                      ...                             \n",
       "21669                      NaT  ...                 NaT     0   \n",
       "54843                      NaT  ...                 NaT     0   \n",
       "77292      2013-01-12 08:50:16  ... 2013-01-12 08:50:16   948   \n",
       "114021     2013-01-12 08:50:18  ... 2013-01-12 08:50:18   947   \n",
       "146670     2013-01-12 08:50:21  ... 2013-01-12 08:50:21   946   \n",
       "242171     2013-01-12 08:50:24  ... 2013-01-12 08:50:24   953   \n",
       "57157      2013-01-12 08:50:26  ... 2013-01-12 08:50:27   953   \n",
       "240201     2013-01-12 08:50:29  ... 2013-01-12 08:50:29   946   \n",
       "210686     2013-01-12 08:50:33  ... 2013-01-12 08:50:33   946   \n",
       "98804      2013-01-12 08:50:59  ... 2013-01-12 08:51:03   812   \n",
       "\n",
       "                         time7 site8               time8 site9  \\\n",
       "session_id                                                       \n",
       "21669                      NaT     0                 NaT     0   \n",
       "54843                      NaT     0                 NaT     0   \n",
       "77292      2013-01-12 08:50:16   784 2013-01-12 08:50:16   949   \n",
       "114021     2013-01-12 08:50:19   945 2013-01-12 08:50:19   946   \n",
       "146670     2013-01-12 08:50:21   951 2013-01-12 08:50:22   946   \n",
       "242171     2013-01-12 08:50:24   955 2013-01-12 08:50:24   946   \n",
       "57157      2013-01-12 08:50:27   946 2013-01-12 08:50:27   953   \n",
       "240201     2013-01-12 08:50:30   956 2013-01-12 08:50:30   957   \n",
       "210686     2013-01-12 08:50:34   946 2013-01-12 08:50:35   946   \n",
       "98804      2013-01-12 08:51:03   982 2013-01-12 08:51:03    52   \n",
       "\n",
       "                         time9 site10              time10 target  \n",
       "session_id                                                        \n",
       "21669                      NaT      0                 NaT      0  \n",
       "54843                      NaT      0                 NaT      0  \n",
       "77292      2013-01-12 08:50:17    946 2013-01-12 08:50:17      0  \n",
       "114021     2013-01-12 08:50:19    946 2013-01-12 08:50:20      0  \n",
       "146670     2013-01-12 08:50:22    947 2013-01-12 08:50:22      0  \n",
       "242171     2013-01-12 08:50:25    947 2013-01-12 08:50:25      0  \n",
       "57157      2013-01-12 08:50:28   1033 2013-01-12 08:50:28      0  \n",
       "240201     2013-01-12 08:50:31    956 2013-01-12 08:50:31      0  \n",
       "210686     2013-01-12 08:50:36    948 2013-01-12 08:50:36      0  \n",
       "98804      2013-01-12 08:51:03     52 2013-01-12 08:51:04      0  \n",
       "\n",
       "[10 rows x 21 columns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>seconds</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>253561.000000</td>\n",
       "      <td>253561.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.009059</td>\n",
       "      <td>138.494370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.094747</td>\n",
       "      <td>295.781742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>7.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>27.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>113.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1800.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              target        seconds\n",
       "count  253561.000000  253561.000000\n",
       "mean        0.009059     138.494370\n",
       "std         0.094747     295.781742\n",
       "min         0.000000       0.000000\n",
       "25%         0.000000       7.000000\n",
       "50%         0.000000      27.000000\n",
       "75%         0.000000     113.000000\n",
       "max         1.000000    1800.000000"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "time_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "#need to create alice and non-alice sessions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(pandas.core.frame.DataFrame, 253561)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(time_df), len(time_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>seconds</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>251175</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:25:10</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>437.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196388</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172448</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>2013-02-12 16:33:46</td>\n",
       "      <td>53.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70129</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:33:50</td>\n",
       "      <td>2013-02-12 16:33:53</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206254</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:33:55</td>\n",
       "      <td>2013-02-12 16:34:25</td>\n",
       "      <td>30.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            target                 min                 max  seconds\n",
       "session_id                                                         \n",
       "251175           1 2013-02-12 16:25:10 2013-02-12 16:32:27    437.0\n",
       "196388           1 2013-02-12 16:32:27 2013-02-12 16:32:53     26.0\n",
       "172448           1 2013-02-12 16:32:53 2013-02-12 16:33:46     53.0\n",
       "70129            1 2013-02-12 16:33:50 2013-02-12 16:33:53      3.0\n",
       "206254           1 2013-02-12 16:33:55 2013-02-12 16:34:25     30.0"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df = time_df[time_df['target']==1]\n",
    "alice_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>seconds</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>251175</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:25:10</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>437.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196388</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172448</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>2013-02-12 16:33:46</td>\n",
       "      <td>53.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70129</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:33:50</td>\n",
       "      <td>2013-02-12 16:33:53</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206254</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:33:55</td>\n",
       "      <td>2013-02-12 16:34:25</td>\n",
       "      <td>30.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            target                 min                 max  seconds\n",
       "session_id                                                         \n",
       "251175           1 2013-02-12 16:25:10 2013-02-12 16:32:27    437.0\n",
       "196388           1 2013-02-12 16:32:27 2013-02-12 16:32:53     26.0\n",
       "172448           1 2013-02-12 16:32:53 2013-02-12 16:33:46     53.0\n",
       "70129            1 2013-02-12 16:33:50 2013-02-12 16:33:53      3.0\n",
       "206254           1 2013-02-12 16:33:55 2013-02-12 16:34:25     30.0"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\pandas\\core\\frame.py:4025: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  return super(DataFrame, self).rename(**kwargs)\n"
     ]
    }
   ],
   "source": [
    "alice_df.rename(columns={'seconds':'sec'}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>sec</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>251175</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:25:10</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>437.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196388</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172448</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>2013-02-12 16:33:46</td>\n",
       "      <td>53.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70129</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:33:50</td>\n",
       "      <td>2013-02-12 16:33:53</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206254</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:33:55</td>\n",
       "      <td>2013-02-12 16:34:25</td>\n",
       "      <td>30.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            target                 min                 max    sec\n",
       "session_id                                                       \n",
       "251175           1 2013-02-12 16:25:10 2013-02-12 16:32:27  437.0\n",
       "196388           1 2013-02-12 16:32:27 2013-02-12 16:32:53   26.0\n",
       "172448           1 2013-02-12 16:32:53 2013-02-12 16:33:46   53.0\n",
       "70129            1 2013-02-12 16:33:50 2013-02-12 16:33:53    3.0\n",
       "206254           1 2013-02-12 16:33:55 2013-02-12 16:34:25   30.0"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    }
   ],
   "source": [
    "alice_df['minutes'] = alice_df['sec']/60"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>sec</th>\n",
       "      <th>minutes</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>251175</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:25:10</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>437.0</td>\n",
       "      <td>7.283333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196388</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:32:27</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0.433333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172448</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:32:53</td>\n",
       "      <td>2013-02-12 16:33:46</td>\n",
       "      <td>53.0</td>\n",
       "      <td>0.883333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70129</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:33:50</td>\n",
       "      <td>2013-02-12 16:33:53</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.050000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206254</th>\n",
       "      <td>1</td>\n",
       "      <td>2013-02-12 16:33:55</td>\n",
       "      <td>2013-02-12 16:34:25</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            target                 min                 max    sec   minutes\n",
       "session_id                                                                 \n",
       "251175           1 2013-02-12 16:25:10 2013-02-12 16:32:27  437.0  7.283333\n",
       "196388           1 2013-02-12 16:32:27 2013-02-12 16:32:53   26.0  0.433333\n",
       "172448           1 2013-02-12 16:32:53 2013-02-12 16:33:46   53.0  0.883333\n",
       "70129            1 2013-02-12 16:33:50 2013-02-12 16:33:53    3.0  0.050000\n",
       "206254           1 2013-02-12 16:33:55 2013-02-12 16:34:25   30.0  0.500000"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>seconds</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>1786.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            target                 min                 max  seconds\n",
       "session_id                                                         \n",
       "21669            0 2013-01-12 08:05:57 2013-01-12 08:05:57      0.0\n",
       "54843            0 2013-01-12 08:37:23 2013-01-12 09:07:09   1786.0\n",
       "77292            0 2013-01-12 08:50:13 2013-01-12 08:50:17      4.0\n",
       "114021           0 2013-01-12 08:50:17 2013-01-12 08:50:20      3.0\n",
       "146670           0 2013-01-12 08:50:20 2013-01-12 08:50:22      2.0"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "other_users = time_df[time_df['target']==0]\n",
    "other_users.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>seconds</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>1786.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            target                 min                 max  seconds\n",
       "session_id                                                         \n",
       "21669            0 2013-01-12 08:05:57 2013-01-12 08:05:57      0.0\n",
       "54843            0 2013-01-12 08:37:23 2013-01-12 09:07:09   1786.0\n",
       "77292            0 2013-01-12 08:50:13 2013-01-12 08:50:17      4.0\n",
       "114021           0 2013-01-12 08:50:17 2013-01-12 08:50:20      3.0\n",
       "146670           0 2013-01-12 08:50:20 2013-01-12 08:50:22      2.0"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "other_users.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>target</th>\n",
       "      <th>min</th>\n",
       "      <th>max</th>\n",
       "      <th>seconds</th>\n",
       "      <th>minutes</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>1786.0</td>\n",
       "      <td>29.766667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.066667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.050000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>0</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.033333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            target                 min                 max  seconds    minutes\n",
       "session_id                                                                    \n",
       "21669            0 2013-01-12 08:05:57 2013-01-12 08:05:57      0.0   0.000000\n",
       "54843            0 2013-01-12 08:37:23 2013-01-12 09:07:09   1786.0  29.766667\n",
       "77292            0 2013-01-12 08:50:13 2013-01-12 08:50:17      4.0   0.066667\n",
       "114021           0 2013-01-12 08:50:17 2013-01-12 08:50:20      3.0   0.050000\n",
       "146670           0 2013-01-12 08:50:20 2013-01-12 08:50:22      2.0   0.033333"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "other_users['minutes'] = other_users['seconds']/60\n",
    "other_users.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "avg session for alice is : 0.8716078943549493\n"
     ]
    }
   ],
   "source": [
    "print(\"avg session for alice is : {}\".format( sum(alice_df['minutes'])/ len(alice_df['minutes'])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "avg session for alice is : 0.8716078943549558\n"
     ]
    }
   ],
   "source": [
    "print( \"avg session for alice is : {}\".format(np.mean((alice_df['minutes'])) ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "avg session for other users is 2.3213728720917524\n"
     ]
    }
   ],
   "source": [
    "print( \"avg session for other users is {}\".format(sum(other_users['minutes'] / len(other_users['minutes']))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "avg session for other users is 2.321372872092036\n"
     ]
    }
   ],
   "source": [
    "print( \"avg session for other users is {}\".format( np.mean(other_users['minutes'])))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "on avg,  alice's session is shorter than other users - TRUE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "29.383333333333333"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.describe().loc['max']['minutes']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "def approx_same(a,b):\n",
    "    return abs(a-b)/ max(a,b) <= 0.05\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "more than 1% of all sessions in the dataset belong to Alice"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2297, 5)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(253561, 21)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.905896411514389"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(2297*100)/253561"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "we have < 1% of sessions belonging to alice"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "maximum durations of Alice's and other users' sessions are approximately the same?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "approx_same(alice_df.describe().loc['max']['minutes'], other_users.describe().loc['max']['minutes'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "29.383333333333333"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.describe().loc['max']['minutes']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "30.0"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "other_users.describe().loc['max']['minutes']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "minimum durations of Alice's and other users' sessions are approximately the same?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\ipykernel_launcher.py:2: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "approx_same(alice_df.describe().loc['min']['minutes'], other_users.describe().loc['min']['minutes'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.describe().loc['min']['minutes']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "other_users.describe().loc['min']['minutes']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "TRUE - MAX & MIN DURATIONS ARE APPROX SAME!!!!!!!!!!!!!!!!!!!!!!!!!!!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "standard deviation of Alice's sessions duration is approximately the same as for non-Alice's sessions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "approx_same(alice_df.describe().loc['std']['minutes'], other_users.describe().loc['std']['minutes'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2.555150234212393"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.describe().loc['std']['minutes']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.944225295308388"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "other_users.describe().loc['std']['minutes']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "less than a quarter of Alice's sessions are greater than or equal to 40 seconds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['target', 'min', 'max', 'sec', 'minutes'], dtype='object')"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alice_df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2297"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(alice_df.sec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "554"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(alice_df[alice_df['sec']>=40])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "less than a quarter of alice sessions are greater than 40seconds ,  25% =24.11841532433609 % , so it is True\n"
     ]
    }
   ],
   "source": [
    "print(\"less than a quarter of alice sessions are greater than 40seconds ,  25% ={0} % , so it is {1}\".format(  (554*100/2297) , (554*100/2297)<25 ))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## train + test = full_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Our target variable\n",
    "y_train = train_df['target']\n",
    "\n",
    "# United dataframe of the initial data \n",
    "full_df = pd.concat([train_df.drop('target', axis=1), test_df])\n",
    "\n",
    "# Index to split the training and test data sets\n",
    "idx_split = train_df.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>site2</th>\n",
       "      <th>site3</th>\n",
       "      <th>site4</th>\n",
       "      <th>site5</th>\n",
       "      <th>site6</th>\n",
       "      <th>site7</th>\n",
       "      <th>site8</th>\n",
       "      <th>site9</th>\n",
       "      <th>site10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>55</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>55</td>\n",
       "      <td>56</td>\n",
       "      <td>55</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>946</td>\n",
       "      <td>951</td>\n",
       "      <td>946</td>\n",
       "      <td>946</td>\n",
       "      <td>945</td>\n",
       "      <td>948</td>\n",
       "      <td>784</td>\n",
       "      <td>949</td>\n",
       "      <td>946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>948</td>\n",
       "      <td>949</td>\n",
       "      <td>948</td>\n",
       "      <td>945</td>\n",
       "      <td>946</td>\n",
       "      <td>947</td>\n",
       "      <td>945</td>\n",
       "      <td>946</td>\n",
       "      <td>946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>950</td>\n",
       "      <td>948</td>\n",
       "      <td>947</td>\n",
       "      <td>950</td>\n",
       "      <td>952</td>\n",
       "      <td>946</td>\n",
       "      <td>951</td>\n",
       "      <td>946</td>\n",
       "      <td>947</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1  site2  site3  site4  site5  site6  site7  site8  site9  \\\n",
       "session_id                                                                  \n",
       "21669          56     55      0      0      0      0      0      0      0   \n",
       "54843          56     55     56     55      0      0      0      0      0   \n",
       "77292         946    946    951    946    946    945    948    784    949   \n",
       "114021        945    948    949    948    945    946    947    945    946   \n",
       "146670        947    950    948    947    950    952    946    951    946   \n",
       "\n",
       "            site10  \n",
       "session_id          \n",
       "21669            0  \n",
       "54843            0  \n",
       "77292          946  \n",
       "114021         946  \n",
       "146670         947  "
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_sites = full_df[sites]\n",
    "full_sites.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "sites_flatten = full_sites.values.flatten()\n",
    "\n",
    "# and the matrix we are looking for \n",
    "# (make sure you understand which of the `csr_matrix` constructors is used here)\n",
    "# a further toy example will help you with it\n",
    "full_sites_sparse = csr_matrix(([1] * sites_flatten.shape[0],\n",
    "                                sites_flatten,\n",
    "                                range(0, sites_flatten.shape[0]  + 10, 10)))[:, 1:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2047      2\n",
       "35985     2\n",
       "60317     2\n",
       "62366     2\n",
       "64415     2\n",
       "66464     2\n",
       "68513     2\n",
       "70562     2\n",
       "72611     2\n",
       "74660     2\n",
       "76709     2\n",
       "78758     2\n",
       "80807     2\n",
       "1152      2\n",
       "3201      2\n",
       "5250      2\n",
       "7299      2\n",
       "9348      2\n",
       "11397     2\n",
       "13446     2\n",
       "15495     2\n",
       "17544     2\n",
       "19593     2\n",
       "21642     2\n",
       "23691     2\n",
       "25740     2\n",
       "27789     2\n",
       "29838     2\n",
       "31887     2\n",
       "58268     2\n",
       "         ..\n",
       "112458    1\n",
       "110409    1\n",
       "108360    1\n",
       "122695    1\n",
       "120646    1\n",
       "118597    1\n",
       "116548    1\n",
       "87894     1\n",
       "131479    1\n",
       "172419    1\n",
       "186764    1\n",
       "170372    1\n",
       "168325    1\n",
       "166278    1\n",
       "164231    1\n",
       "194952    1\n",
       "192905    1\n",
       "190858    1\n",
       "188811    1\n",
       "184717    1\n",
       "133526    1\n",
       "182670    1\n",
       "180623    1\n",
       "145808    1\n",
       "143761    1\n",
       "141714    1\n",
       "139667    1\n",
       "137620    1\n",
       "135573    1\n",
       "124739    1\n",
       "Name: session_id, Length: 253561, dtype: int64"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_sites.index.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>site2</th>\n",
       "      <th>site3</th>\n",
       "      <th>site4</th>\n",
       "      <th>site5</th>\n",
       "      <th>site6</th>\n",
       "      <th>site7</th>\n",
       "      <th>site8</th>\n",
       "      <th>site9</th>\n",
       "      <th>site10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>950</td>\n",
       "      <td>948</td>\n",
       "      <td>947</td>\n",
       "      <td>950</td>\n",
       "      <td>952</td>\n",
       "      <td>946</td>\n",
       "      <td>951</td>\n",
       "      <td>946</td>\n",
       "      <td>947</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1  site2  site3  site4  site5  site6  site7  site8  site9  \\\n",
       "session_id                                                                  \n",
       "146670        947    950    948    947    950    952    946    951    946   \n",
       "\n",
       "            site10  \n",
       "session_id          \n",
       "146670         947  "
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_sites[full_sites.index == 146670]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(336358, 48371)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_sites_sparse.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1866898 elements * 8 bytes = 14935184 bytes\n",
      "sparse_matrix_size = 7467592 bytes\n"
     ]
    }
   ],
   "source": [
    "print('{0} elements * {1} bytes = {2} bytes'.format(full_sites_sparse.count_nonzero(), 8, \n",
    "                                                    full_sites_sparse.count_nonzero() * 8))\n",
    "# Or just like this:\n",
    "print('sparse_matrix_size = {0} bytes'.format(full_sites_sparse.data.nbytes))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = [1] * 9\n",
    "\n",
    "# To do this, you need to correctly distribute the ones in cells\n",
    "# Indices - website ids, i.e. columns of a new matrix. We will sum ones up grouping them by sessions (ids)\n",
    "indices = [1, 0, 0, 1, 3, 1, 2, 3, 4]\n",
    "\n",
    "# Indices for the division into rows (sessions)\n",
    "# For example, line 0 is the elements between the indices [0; 3) - the rightmost value is not included\n",
    "# Line 1 is the elements between the indices [3; 6)\n",
    "# Line 2 is the elements between the indices [6; 9) \n",
    "indptr = [0, 3, 6, 9]\n",
    "\n",
    "# Aggregate these three variables into a tuple and compose a matrix\n",
    "# To display this matrix on the screen transform it into the usual \"dense\" matrix\n",
    "dense_m = csr_matrix((data, indices, indptr)).todense()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "dense_m = pd.DataFrame(dense_m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0  1  2  3  4\n",
       "0  2  1  0  0  0\n",
       "1  0  2  0  1  0\n",
       "2  0  0  1  1  1"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dense_m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "dense_m = dense_m.drop(columns=[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   1  2  3  4\n",
       "0  1  0  0  0\n",
       "1  2  0  1  0\n",
       "2  0  1  1  1"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dense_m"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "non_zero elements = 6, zero-elem = 6, total = 12\n",
    "\n",
    "sparcity = 1 - (nonzero/total) = 1 - (6/12) = 1- 0.5 = 0.5, 50%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_auc_lr_valid(X, y, C=1.0, seed=17, ratio = 0.9):\n",
    "    # Split the data into the training and validation sets\n",
    "    idx = int(round(X.shape[0] * ratio))\n",
    "    # Classifier training\n",
    "    lr = LogisticRegression(C=C, random_state=seed, solver='liblinear').fit(X[:idx, :], y[:idx])\n",
    "    # Prediction for validation set\n",
    "    y_pred = lr.predict_proba(X[idx:, :])[:, 1]\n",
    "    # Calculate the quality\n",
    "    score = roc_auc_score(y[idx:], y_pred)\n",
    "    \n",
    "    return score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9195245587148385\n",
      "Wall time: 12.7 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Select the training set from the united dataframe (where we have the answers)\n",
    "X_train = full_sites_sparse[:idx_split, :]\n",
    "\n",
    "# Calculate metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9195245587148385\n",
      "Wall time: 12.3 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Select the training set from the united dataframe (where we have the answers)\n",
    "X_train = full_sites_sparse[:idx_split, :]\n",
    "\n",
    "# Calculate metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function for writing predictions to a file\n",
    "def write_to_submission_file(predicted_labels, out_file,\n",
    "                             target='target', index_label=\"session_id\"):\n",
    "    predicted_df = pd.DataFrame(predicted_labels,\n",
    "                                index = np.arange(1, predicted_labels.shape[0] + 1),\n",
    "                                columns=[target])\n",
    "    predicted_df.to_csv(out_file, index_label=index_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train the model on the whole training data set\n",
    "# Use random_state=17 for repeatability\n",
    "# Parameter C=1 by default, but here we set it explicitly\n",
    "lr = LogisticRegression(C=1.0, random_state=17, solver='liblinear').fit(X_train, y_train)\n",
    "\n",
    "# Make a prediction for test data set\n",
    "X_test = full_sites_sparse[idx_split:,:]\n",
    "y_test = lr.predict_proba(X_test)[:, 1]\n",
    "\n",
    "# Write it to the file which could be submitted\n",
    "write_to_submission_file(y_test, 'baseline_1.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Adding start_month"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>site6</th>\n",
       "      <th>time6</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 09:07:07</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:14</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>784</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>952</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "21669          56 2013-01-12 08:05:57     55 2013-01-12 08:05:57      0   \n",
       "54843          56 2013-01-12 08:37:23     55 2013-01-12 08:37:23     56   \n",
       "77292         946 2013-01-12 08:50:13    946 2013-01-12 08:50:14    951   \n",
       "114021        945 2013-01-12 08:50:17    948 2013-01-12 08:50:17    949   \n",
       "146670        947 2013-01-12 08:50:20    950 2013-01-12 08:50:20    948   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843      2013-01-12 09:07:07     55 2013-01-12 09:07:09      0   \n",
       "77292      2013-01-12 08:50:15    946 2013-01-12 08:50:15    946   \n",
       "114021     2013-01-12 08:50:18    948 2013-01-12 08:50:18    945   \n",
       "146670     2013-01-12 08:50:20    947 2013-01-12 08:50:21    950   \n",
       "\n",
       "                         time5  site6               time6  site7  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843                      NaT      0                 NaT      0   \n",
       "77292      2013-01-12 08:50:16    945 2013-01-12 08:50:16    948   \n",
       "114021     2013-01-12 08:50:18    946 2013-01-12 08:50:18    947   \n",
       "146670     2013-01-12 08:50:21    952 2013-01-12 08:50:21    946   \n",
       "\n",
       "                         time7  site8               time8  site9  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843                      NaT      0                 NaT      0   \n",
       "77292      2013-01-12 08:50:16    784 2013-01-12 08:50:16    949   \n",
       "114021     2013-01-12 08:50:19    945 2013-01-12 08:50:19    946   \n",
       "146670     2013-01-12 08:50:21    951 2013-01-12 08:50:22    946   \n",
       "\n",
       "                         time9  site10              time10  \n",
       "session_id                                                  \n",
       "21669                      NaT       0                 NaT  \n",
       "54843                      NaT       0                 NaT  \n",
       "77292      2013-01-12 08:50:17     946 2013-01-12 08:50:17  \n",
       "114021     2013-01-12 08:50:19     946 2013-01-12 08:50:20  \n",
       "146670     2013-01-12 08:50:22     947 2013-01-12 08:50:22  "
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>site6</th>\n",
       "      <th>time6</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>21</td>\n",
       "      <td>2014-05-05 15:28:30</td>\n",
       "      <td>42279</td>\n",
       "      <td>2014-05-05 15:28:31</td>\n",
       "      <td>76</td>\n",
       "      <td>2014-05-05 15:28:31</td>\n",
       "      <td>21</td>\n",
       "      <td>2014-05-05 15:28:31</td>\n",
       "      <td>229</td>\n",
       "      <td>2014-05-05 15:28:31</td>\n",
       "      <td>21</td>\n",
       "      <td>2014-05-05 15:28:32</td>\n",
       "      <td>234</td>\n",
       "      <td>2014-05-05 15:28:32</td>\n",
       "      <td>42279</td>\n",
       "      <td>2014-05-05 15:28:32</td>\n",
       "      <td>76</td>\n",
       "      <td>2014-05-05 15:28:33</td>\n",
       "      <td>42279</td>\n",
       "      <td>2014-05-05 15:28:33</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "21669          56 2013-01-12 08:05:57     55 2013-01-12 08:05:57      0   \n",
       "21669          21 2014-05-05 15:28:30  42279 2014-05-05 15:28:31     76   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "21669      2014-05-05 15:28:31     21 2014-05-05 15:28:31    229   \n",
       "\n",
       "                         time5  site6               time6  site7  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "21669      2014-05-05 15:28:31     21 2014-05-05 15:28:32    234   \n",
       "\n",
       "                         time7  site8               time8  site9  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "21669      2014-05-05 15:28:32  42279 2014-05-05 15:28:32     76   \n",
       "\n",
       "                         time9  site10              time10  \n",
       "session_id                                                  \n",
       "21669                      NaT       0                 NaT  \n",
       "21669      2014-05-05 15:28:33   42279 2014-05-05 15:28:33  "
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_df[full_df.index == 21669]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_new_feat = pd.DataFrame(index=full_df.index)\n",
    "\n",
    "# Add start_month feature\n",
    "full_new_feat['start_month'] = full_df['time1'].apply(lambda ts: \n",
    "                                                      100 * ts.year + ts.month).astype('float64')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot the graph of the number of Alice sessions versus the new feature, start_month. Choose the correct statemen"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>start_month</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            start_month\n",
       "session_id             \n",
       "21669          201301.0\n",
       "54843          201301.0\n",
       "77292          201301.0\n",
       "114021         201301.0\n",
       "146670         201301.0"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_new_feat.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_new_feat['start_month'] = full_new_feat['start_month'].astype('category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0xb1f8cb0>"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 580,
       "width": 1188
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,10))\n",
    "full_new_feat['start_month'].value_counts().sort_index().plot(kind='barh')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "answer - From the beginning of 2013 to mid-2014, the number of Alice's sessions per month increased"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### calculating ROC auc without scaling for start_month"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7508354860175162\n"
     ]
    }
   ],
   "source": [
    "# Add the new feature to the sparse matrix\n",
    "tmp = full_new_feat[['start_month']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### calculating ROC auc with scaling for start_month"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype category were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype category were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9196827643967169\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_new_feat[['start_month']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Adding new feature - n_unique_visits to both train & test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>...</th>\n",
       "      <th>time6</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 09:07:07</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:14</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>784</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "21669          56 2013-01-12 08:05:57     55 2013-01-12 08:05:57      0   \n",
       "54843          56 2013-01-12 08:37:23     55 2013-01-12 08:37:23     56   \n",
       "77292         946 2013-01-12 08:50:13    946 2013-01-12 08:50:14    951   \n",
       "114021        945 2013-01-12 08:50:17    948 2013-01-12 08:50:17    949   \n",
       "146670        947 2013-01-12 08:50:20    950 2013-01-12 08:50:20    948   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843      2013-01-12 09:07:07     55 2013-01-12 09:07:09      0   \n",
       "77292      2013-01-12 08:50:15    946 2013-01-12 08:50:15    946   \n",
       "114021     2013-01-12 08:50:18    948 2013-01-12 08:50:18    945   \n",
       "146670     2013-01-12 08:50:20    947 2013-01-12 08:50:21    950   \n",
       "\n",
       "                         time5  ...               time6 site7  \\\n",
       "session_id                      ...                             \n",
       "21669                      NaT  ...                 NaT     0   \n",
       "54843                      NaT  ...                 NaT     0   \n",
       "77292      2013-01-12 08:50:16  ... 2013-01-12 08:50:16   948   \n",
       "114021     2013-01-12 08:50:18  ... 2013-01-12 08:50:18   947   \n",
       "146670     2013-01-12 08:50:21  ... 2013-01-12 08:50:21   946   \n",
       "\n",
       "                         time7 site8               time8 site9  \\\n",
       "session_id                                                       \n",
       "21669                      NaT     0                 NaT     0   \n",
       "54843                      NaT     0                 NaT     0   \n",
       "77292      2013-01-12 08:50:16   784 2013-01-12 08:50:16   949   \n",
       "114021     2013-01-12 08:50:19   945 2013-01-12 08:50:19   946   \n",
       "146670     2013-01-12 08:50:21   951 2013-01-12 08:50:22   946   \n",
       "\n",
       "                         time9 site10              time10 target  \n",
       "session_id                                                        \n",
       "21669                      NaT      0                 NaT      0  \n",
       "54843                      NaT      0                 NaT      0  \n",
       "77292      2013-01-12 08:50:17    946 2013-01-12 08:50:17      0  \n",
       "114021     2013-01-12 08:50:19    946 2013-01-12 08:50:20      0  \n",
       "146670     2013-01-12 08:50:22    947 2013-01-12 08:50:22      0  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df['start_month']  = train_df['time1'].apply(lambda ts: 100 * ts.year + ts.month).astype('float64')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>...</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "      <th>target</th>\n",
       "      <th>start_month</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 09:07:07</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:14</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>...</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>784</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>...</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>...</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "21669          56 2013-01-12 08:05:57     55 2013-01-12 08:05:57      0   \n",
       "54843          56 2013-01-12 08:37:23     55 2013-01-12 08:37:23     56   \n",
       "77292         946 2013-01-12 08:50:13    946 2013-01-12 08:50:14    951   \n",
       "114021        945 2013-01-12 08:50:17    948 2013-01-12 08:50:17    949   \n",
       "146670        947 2013-01-12 08:50:20    950 2013-01-12 08:50:20    948   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843      2013-01-12 09:07:07     55 2013-01-12 09:07:09      0   \n",
       "77292      2013-01-12 08:50:15    946 2013-01-12 08:50:15    946   \n",
       "114021     2013-01-12 08:50:18    948 2013-01-12 08:50:18    945   \n",
       "146670     2013-01-12 08:50:20    947 2013-01-12 08:50:21    950   \n",
       "\n",
       "                         time5  ...  site7               time7  site8  \\\n",
       "session_id                      ...                                     \n",
       "21669                      NaT  ...      0                 NaT      0   \n",
       "54843                      NaT  ...      0                 NaT      0   \n",
       "77292      2013-01-12 08:50:16  ...    948 2013-01-12 08:50:16    784   \n",
       "114021     2013-01-12 08:50:18  ...    947 2013-01-12 08:50:19    945   \n",
       "146670     2013-01-12 08:50:21  ...    946 2013-01-12 08:50:21    951   \n",
       "\n",
       "                         time8  site9               time9  site10  \\\n",
       "session_id                                                          \n",
       "21669                      NaT      0                 NaT       0   \n",
       "54843                      NaT      0                 NaT       0   \n",
       "77292      2013-01-12 08:50:16    949 2013-01-12 08:50:17     946   \n",
       "114021     2013-01-12 08:50:19    946 2013-01-12 08:50:19     946   \n",
       "146670     2013-01-12 08:50:22    946 2013-01-12 08:50:22     947   \n",
       "\n",
       "                        time10  target start_month  \n",
       "session_id                                          \n",
       "21669                      NaT       0    201301.0  \n",
       "54843                      NaT       0    201301.0  \n",
       "77292      2013-01-12 08:50:17       0    201301.0  \n",
       "114021     2013-01-12 08:50:20       0    201301.0  \n",
       "146670     2013-01-12 08:50:22       0    201301.0  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df['n_unique_visits'] = train_df[['site1','site2','site3','site4','site5','site6','site7','site8','site9','site10']].nunique(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>...</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "      <th>target</th>\n",
       "      <th>start_month</th>\n",
       "      <th>n_unique_visits</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 09:07:07</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:14</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>784</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>...</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>0</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "21669          56 2013-01-12 08:05:57     55 2013-01-12 08:05:57      0   \n",
       "54843          56 2013-01-12 08:37:23     55 2013-01-12 08:37:23     56   \n",
       "77292         946 2013-01-12 08:50:13    946 2013-01-12 08:50:14    951   \n",
       "114021        945 2013-01-12 08:50:17    948 2013-01-12 08:50:17    949   \n",
       "146670        947 2013-01-12 08:50:20    950 2013-01-12 08:50:20    948   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843      2013-01-12 09:07:07     55 2013-01-12 09:07:09      0   \n",
       "77292      2013-01-12 08:50:15    946 2013-01-12 08:50:15    946   \n",
       "114021     2013-01-12 08:50:18    948 2013-01-12 08:50:18    945   \n",
       "146670     2013-01-12 08:50:20    947 2013-01-12 08:50:21    950   \n",
       "\n",
       "                         time5  ...               time7 site8  \\\n",
       "session_id                      ...                             \n",
       "21669                      NaT  ...                 NaT     0   \n",
       "54843                      NaT  ...                 NaT     0   \n",
       "77292      2013-01-12 08:50:16  ... 2013-01-12 08:50:16   784   \n",
       "114021     2013-01-12 08:50:18  ... 2013-01-12 08:50:19   945   \n",
       "146670     2013-01-12 08:50:21  ... 2013-01-12 08:50:21   951   \n",
       "\n",
       "                         time8 site9               time9 site10  \\\n",
       "session_id                                                        \n",
       "21669                      NaT     0                 NaT      0   \n",
       "54843                      NaT     0                 NaT      0   \n",
       "77292      2013-01-12 08:50:16   949 2013-01-12 08:50:17    946   \n",
       "114021     2013-01-12 08:50:19   946 2013-01-12 08:50:19    946   \n",
       "146670     2013-01-12 08:50:22   946 2013-01-12 08:50:22    947   \n",
       "\n",
       "                        time10 target  start_month n_unique_visits  \n",
       "session_id                                                          \n",
       "21669                      NaT      0     201301.0               3  \n",
       "54843                      NaT      0     201301.0               3  \n",
       "77292      2013-01-12 08:50:17      0     201301.0               6  \n",
       "114021     2013-01-12 08:50:20      0     201301.0               5  \n",
       "146670     2013-01-12 08:50:22      0     201301.0               6  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " Add to the training set a new feature \"n_unique_sites\" – the number of the unique web-sites in a session. Calculate how the quality on the validation set has changed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df['start_month']  = test_df['time1'].apply(lambda ts: 100 * ts.year + ts.month).astype('float64')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>...</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "      <th>start_month</th>\n",
       "      <th>n_unique_visits</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>29</td>\n",
       "      <td>2014-10-04 11:19:53</td>\n",
       "      <td>35</td>\n",
       "      <td>2014-10-04 11:19:53</td>\n",
       "      <td>22</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>321</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>23</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>...</td>\n",
       "      <td>6730</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>21</td>\n",
       "      <td>2014-10-04 11:19:54</td>\n",
       "      <td>44582</td>\n",
       "      <td>2014-10-04 11:20:00</td>\n",
       "      <td>15336</td>\n",
       "      <td>2014-10-04 11:20:00</td>\n",
       "      <td>201410.0</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:00:28</td>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:00:53</td>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:00:58</td>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:01:06</td>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:01:09</td>\n",
       "      <td>...</td>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:01:23</td>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:01:29</td>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:01:30</td>\n",
       "      <td>782</td>\n",
       "      <td>2014-07-03 11:01:53</td>\n",
       "      <td>201407.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>55</td>\n",
       "      <td>2014-12-05 15:55:12</td>\n",
       "      <td>55</td>\n",
       "      <td>2014-12-05 15:55:13</td>\n",
       "      <td>55</td>\n",
       "      <td>2014-12-05 15:55:14</td>\n",
       "      <td>55</td>\n",
       "      <td>2014-12-05 15:56:15</td>\n",
       "      <td>55</td>\n",
       "      <td>2014-12-05 15:56:16</td>\n",
       "      <td>...</td>\n",
       "      <td>55</td>\n",
       "      <td>2014-12-05 15:56:18</td>\n",
       "      <td>55</td>\n",
       "      <td>2014-12-05 15:56:19</td>\n",
       "      <td>1445</td>\n",
       "      <td>2014-12-05 15:56:33</td>\n",
       "      <td>1445</td>\n",
       "      <td>2014-12-05 15:56:36</td>\n",
       "      <td>201412.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1023</td>\n",
       "      <td>2014-11-04 10:03:19</td>\n",
       "      <td>1022</td>\n",
       "      <td>2014-11-04 10:03:19</td>\n",
       "      <td>50</td>\n",
       "      <td>2014-11-04 10:03:20</td>\n",
       "      <td>222</td>\n",
       "      <td>2014-11-04 10:03:21</td>\n",
       "      <td>202</td>\n",
       "      <td>2014-11-04 10:03:21</td>\n",
       "      <td>...</td>\n",
       "      <td>50</td>\n",
       "      <td>2014-11-04 10:03:22</td>\n",
       "      <td>48</td>\n",
       "      <td>2014-11-04 10:03:22</td>\n",
       "      <td>48</td>\n",
       "      <td>2014-11-04 10:03:23</td>\n",
       "      <td>3374</td>\n",
       "      <td>2014-11-04 10:03:23</td>\n",
       "      <td>201411.0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>301</td>\n",
       "      <td>2014-05-16 15:05:31</td>\n",
       "      <td>301</td>\n",
       "      <td>2014-05-16 15:05:32</td>\n",
       "      <td>301</td>\n",
       "      <td>2014-05-16 15:05:33</td>\n",
       "      <td>66</td>\n",
       "      <td>2014-05-16 15:05:39</td>\n",
       "      <td>67</td>\n",
       "      <td>2014-05-16 15:05:40</td>\n",
       "      <td>...</td>\n",
       "      <td>70</td>\n",
       "      <td>2014-05-16 15:05:40</td>\n",
       "      <td>68</td>\n",
       "      <td>2014-05-16 15:05:40</td>\n",
       "      <td>71</td>\n",
       "      <td>2014-05-16 15:05:40</td>\n",
       "      <td>167</td>\n",
       "      <td>2014-05-16 15:05:44</td>\n",
       "      <td>201405.0</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "1              29 2014-10-04 11:19:53     35 2014-10-04 11:19:53     22   \n",
       "2             782 2014-07-03 11:00:28    782 2014-07-03 11:00:53    782   \n",
       "3              55 2014-12-05 15:55:12     55 2014-12-05 15:55:13     55   \n",
       "4            1023 2014-11-04 10:03:19   1022 2014-11-04 10:03:19     50   \n",
       "5             301 2014-05-16 15:05:31    301 2014-05-16 15:05:32    301   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "1          2014-10-04 11:19:54    321 2014-10-04 11:19:54     23   \n",
       "2          2014-07-03 11:00:58    782 2014-07-03 11:01:06    782   \n",
       "3          2014-12-05 15:55:14     55 2014-12-05 15:56:15     55   \n",
       "4          2014-11-04 10:03:20    222 2014-11-04 10:03:21    202   \n",
       "5          2014-05-16 15:05:33     66 2014-05-16 15:05:39     67   \n",
       "\n",
       "                         time5  ...  site7               time7  site8  \\\n",
       "session_id                      ...                                     \n",
       "1          2014-10-04 11:19:54  ...   6730 2014-10-04 11:19:54     21   \n",
       "2          2014-07-03 11:01:09  ...    782 2014-07-03 11:01:23    782   \n",
       "3          2014-12-05 15:56:16  ...     55 2014-12-05 15:56:18     55   \n",
       "4          2014-11-04 10:03:21  ...     50 2014-11-04 10:03:22     48   \n",
       "5          2014-05-16 15:05:40  ...     70 2014-05-16 15:05:40     68   \n",
       "\n",
       "                         time8  site9               time9  site10  \\\n",
       "session_id                                                          \n",
       "1          2014-10-04 11:19:54  44582 2014-10-04 11:20:00   15336   \n",
       "2          2014-07-03 11:01:29    782 2014-07-03 11:01:30     782   \n",
       "3          2014-12-05 15:56:19   1445 2014-12-05 15:56:33    1445   \n",
       "4          2014-11-04 10:03:22     48 2014-11-04 10:03:23    3374   \n",
       "5          2014-05-16 15:05:40     71 2014-05-16 15:05:40     167   \n",
       "\n",
       "                        time10  start_month n_unique_visits  \n",
       "session_id                                                   \n",
       "1          2014-10-04 11:20:00     201410.0              10  \n",
       "2          2014-07-03 11:01:53     201407.0               1  \n",
       "3          2014-12-05 15:56:36     201412.0               2  \n",
       "4          2014-11-04 10:03:23     201411.0               7  \n",
       "5          2014-05-16 15:05:44     201405.0               8  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df['n_unique_visits'] = test_df[['site1','site2','site3','site4','site5','site6','site7','site8','site9','site10']].nunique(axis=1)\n",
    "test_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(22, 23)"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(test_df.columns), len( train_df.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### concatenating train + test = full_df with new features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Our target variable\n",
    "y_train = train_df['target']\n",
    "\n",
    "# United dataframe of the initial data \n",
    "full_df_with_new_features = pd.concat([train_df.drop('target', axis=1), test_df])\n",
    "\n",
    "# Index to split the training and test data sets\n",
    "idx_split = train_df.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>site2</th>\n",
       "      <th>site3</th>\n",
       "      <th>site4</th>\n",
       "      <th>site5</th>\n",
       "      <th>site6</th>\n",
       "      <th>site7</th>\n",
       "      <th>site8</th>\n",
       "      <th>site9</th>\n",
       "      <th>site10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>55</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>55</td>\n",
       "      <td>56</td>\n",
       "      <td>55</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>946</td>\n",
       "      <td>951</td>\n",
       "      <td>946</td>\n",
       "      <td>946</td>\n",
       "      <td>945</td>\n",
       "      <td>948</td>\n",
       "      <td>784</td>\n",
       "      <td>949</td>\n",
       "      <td>946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>948</td>\n",
       "      <td>949</td>\n",
       "      <td>948</td>\n",
       "      <td>945</td>\n",
       "      <td>946</td>\n",
       "      <td>947</td>\n",
       "      <td>945</td>\n",
       "      <td>946</td>\n",
       "      <td>946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>950</td>\n",
       "      <td>948</td>\n",
       "      <td>947</td>\n",
       "      <td>950</td>\n",
       "      <td>952</td>\n",
       "      <td>946</td>\n",
       "      <td>951</td>\n",
       "      <td>946</td>\n",
       "      <td>947</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1  site2  site3  site4  site5  site6  site7  site8  site9  \\\n",
       "session_id                                                                  \n",
       "21669          56     55      0      0      0      0      0      0      0   \n",
       "54843          56     55     56     55      0      0      0      0      0   \n",
       "77292         946    946    951    946    946    945    948    784    949   \n",
       "114021        945    948    949    948    945    946    947    945    946   \n",
       "146670        947    950    948    947    950    952    946    951    946   \n",
       "\n",
       "            site10  \n",
       "session_id          \n",
       "21669            0  \n",
       "54843            0  \n",
       "77292          946  \n",
       "114021         946  \n",
       "146670         947  "
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_sites = full_df_with_new_features[sites]\n",
    "full_sites.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "sites_flatten = full_sites.values.flatten()\n",
    "\n",
    "# and the matrix we are looking for \n",
    "# (make sure you understand which of the `csr_matrix` constructors is used here)\n",
    "# a further toy example will help you with it\n",
    "full_sites_sparse = csr_matrix(([1] * sites_flatten.shape[0],\n",
    "                                sites_flatten,\n",
    "                                range(0, sites_flatten.shape[0]  + 10, 10)))[:, 1:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### calculating without scaler for start_month"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7508354860175162\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['start_month']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### calculating with scaler for start_month"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9196827643967169\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_df_with_new_features[['start_month']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "0.7508354860175162\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### adding start_hour to full_new_features - no need to repeat adding this feature for train,test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>...</th>\n",
       "      <th>site7</th>\n",
       "      <th>time7</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "      <th>start_month</th>\n",
       "      <th>n_unique_visits</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 09:07:07</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:14</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>...</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>784</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>...</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>...</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "21669          56 2013-01-12 08:05:57     55 2013-01-12 08:05:57      0   \n",
       "54843          56 2013-01-12 08:37:23     55 2013-01-12 08:37:23     56   \n",
       "77292         946 2013-01-12 08:50:13    946 2013-01-12 08:50:14    951   \n",
       "114021        945 2013-01-12 08:50:17    948 2013-01-12 08:50:17    949   \n",
       "146670        947 2013-01-12 08:50:20    950 2013-01-12 08:50:20    948   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843      2013-01-12 09:07:07     55 2013-01-12 09:07:09      0   \n",
       "77292      2013-01-12 08:50:15    946 2013-01-12 08:50:15    946   \n",
       "114021     2013-01-12 08:50:18    948 2013-01-12 08:50:18    945   \n",
       "146670     2013-01-12 08:50:20    947 2013-01-12 08:50:21    950   \n",
       "\n",
       "                         time5  ...  site7               time7  site8  \\\n",
       "session_id                      ...                                     \n",
       "21669                      NaT  ...      0                 NaT      0   \n",
       "54843                      NaT  ...      0                 NaT      0   \n",
       "77292      2013-01-12 08:50:16  ...    948 2013-01-12 08:50:16    784   \n",
       "114021     2013-01-12 08:50:18  ...    947 2013-01-12 08:50:19    945   \n",
       "146670     2013-01-12 08:50:21  ...    946 2013-01-12 08:50:21    951   \n",
       "\n",
       "                         time8  site9               time9  site10  \\\n",
       "session_id                                                          \n",
       "21669                      NaT      0                 NaT       0   \n",
       "54843                      NaT      0                 NaT       0   \n",
       "77292      2013-01-12 08:50:16    949 2013-01-12 08:50:17     946   \n",
       "114021     2013-01-12 08:50:19    946 2013-01-12 08:50:19     946   \n",
       "146670     2013-01-12 08:50:22    946 2013-01-12 08:50:22     947   \n",
       "\n",
       "                        time10  start_month n_unique_visits  \n",
       "session_id                                                   \n",
       "21669                      NaT     201301.0               3  \n",
       "54843                      NaT     201301.0               3  \n",
       "77292      2013-01-12 08:50:17     201301.0               6  \n",
       "114021     2013-01-12 08:50:20     201301.0               5  \n",
       "146670     2013-01-12 08:50:22     201301.0               6  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_df_with_new_features.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_df_with_new_features['start_hour'] = full_df_with_new_features['time1'].apply(lambda time: time.hour)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Timestamp('2013-10-12 09:53:49')"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = full_df_with_new_features['time1'].iloc[21669]\n",
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "int"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(x.hour)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Cannot add integral value to Timestamp without freq.",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-104-ac876cc78eef>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mmy_own_exper\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'start_'\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mfull_df_with_new_features\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'time1'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mapply\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;32mlambda\u001b[0m \u001b[0mtime\u001b[0m\u001b[1;33m:\u001b[0m \u001b[0mtime\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mhour\u001b[0m \u001b[1;33m+\u001b[0m \u001b[0mtime\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32mc:\\python3\\lib\\site-packages\\pandas\\core\\series.py\u001b[0m in \u001b[0;36mapply\u001b[1;34m(self, func, convert_dtype, args, **kwds)\u001b[0m\n\u001b[0;32m   3589\u001b[0m             \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3590\u001b[0m                 \u001b[0mvalues\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mobject\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 3591\u001b[1;33m                 \u001b[0mmapped\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlib\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmap_infer\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mvalues\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mf\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mconvert\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mconvert_dtype\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   3592\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3593\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmapped\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mand\u001b[0m \u001b[0misinstance\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmapped\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mSeries\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mpandas\\_libs\\lib.pyx\u001b[0m in \u001b[0;36mpandas._libs.lib.map_infer\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;32m<ipython-input-104-ac876cc78eef>\u001b[0m in \u001b[0;36m<lambda>\u001b[1;34m(time)\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mmy_own_exper\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'start_'\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mfull_df_with_new_features\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'time1'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mapply\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;32mlambda\u001b[0m \u001b[0mtime\u001b[0m\u001b[1;33m:\u001b[0m \u001b[0mtime\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mhour\u001b[0m \u001b[1;33m+\u001b[0m \u001b[0mtime\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32mpandas\\_libs\\tslibs\\timestamps.pyx\u001b[0m in \u001b[0;36mpandas._libs.tslibs.timestamps.Timestamp.__radd__\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;32mpandas\\_libs\\tslibs\\timestamps.pyx\u001b[0m in \u001b[0;36mpandas._libs.tslibs.timestamps._Timestamp.__add__\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;31mValueError\u001b[0m: Cannot add integral value to Timestamp without freq."
     ]
    }
   ],
   "source": [
    "my_own_exper['start_'] = full_df_with_new_features['time1'].apply(lambda time: time.hour + time)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_df_with_new_features['morning'] = full_df_with_new_features['start_hour'].apply(lambda x: 1 if x <= 11 else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>site1</th>\n",
       "      <th>time1</th>\n",
       "      <th>site2</th>\n",
       "      <th>time2</th>\n",
       "      <th>site3</th>\n",
       "      <th>time3</th>\n",
       "      <th>site4</th>\n",
       "      <th>time4</th>\n",
       "      <th>site5</th>\n",
       "      <th>time5</th>\n",
       "      <th>...</th>\n",
       "      <th>site8</th>\n",
       "      <th>time8</th>\n",
       "      <th>site9</th>\n",
       "      <th>time9</th>\n",
       "      <th>site10</th>\n",
       "      <th>time10</th>\n",
       "      <th>start_month</th>\n",
       "      <th>n_unique_visits</th>\n",
       "      <th>start_hour</th>\n",
       "      <th>morning</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21669</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:05:57</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54843</th>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 08:37:23</td>\n",
       "      <td>56</td>\n",
       "      <td>2013-01-12 09:07:07</td>\n",
       "      <td>55</td>\n",
       "      <td>2013-01-12 09:07:09</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>0</td>\n",
       "      <td>NaT</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77292</th>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:13</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:14</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:15</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>...</td>\n",
       "      <td>784</td>\n",
       "      <td>2013-01-12 08:50:16</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114021</th>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:17</td>\n",
       "      <td>949</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:18</td>\n",
       "      <td>...</td>\n",
       "      <td>945</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:19</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>5</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146670</th>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>948</td>\n",
       "      <td>2013-01-12 08:50:20</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>950</td>\n",
       "      <td>2013-01-12 08:50:21</td>\n",
       "      <td>...</td>\n",
       "      <td>951</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>946</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>947</td>\n",
       "      <td>2013-01-12 08:50:22</td>\n",
       "      <td>201301.0</td>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            site1               time1  site2               time2  site3  \\\n",
       "session_id                                                                \n",
       "21669          56 2013-01-12 08:05:57     55 2013-01-12 08:05:57      0   \n",
       "54843          56 2013-01-12 08:37:23     55 2013-01-12 08:37:23     56   \n",
       "77292         946 2013-01-12 08:50:13    946 2013-01-12 08:50:14    951   \n",
       "114021        945 2013-01-12 08:50:17    948 2013-01-12 08:50:17    949   \n",
       "146670        947 2013-01-12 08:50:20    950 2013-01-12 08:50:20    948   \n",
       "\n",
       "                         time3  site4               time4  site5  \\\n",
       "session_id                                                         \n",
       "21669                      NaT      0                 NaT      0   \n",
       "54843      2013-01-12 09:07:07     55 2013-01-12 09:07:09      0   \n",
       "77292      2013-01-12 08:50:15    946 2013-01-12 08:50:15    946   \n",
       "114021     2013-01-12 08:50:18    948 2013-01-12 08:50:18    945   \n",
       "146670     2013-01-12 08:50:20    947 2013-01-12 08:50:21    950   \n",
       "\n",
       "                         time5  ...  site8               time8  site9  \\\n",
       "session_id                      ...                                     \n",
       "21669                      NaT  ...      0                 NaT      0   \n",
       "54843                      NaT  ...      0                 NaT      0   \n",
       "77292      2013-01-12 08:50:16  ...    784 2013-01-12 08:50:16    949   \n",
       "114021     2013-01-12 08:50:18  ...    945 2013-01-12 08:50:19    946   \n",
       "146670     2013-01-12 08:50:21  ...    951 2013-01-12 08:50:22    946   \n",
       "\n",
       "                         time9  site10              time10  start_month  \\\n",
       "session_id                                                                \n",
       "21669                      NaT       0                 NaT     201301.0   \n",
       "54843                      NaT       0                 NaT     201301.0   \n",
       "77292      2013-01-12 08:50:17     946 2013-01-12 08:50:17     201301.0   \n",
       "114021     2013-01-12 08:50:19     946 2013-01-12 08:50:20     201301.0   \n",
       "146670     2013-01-12 08:50:22     947 2013-01-12 08:50:22     201301.0   \n",
       "\n",
       "           n_unique_visits  start_hour morning  \n",
       "session_id                                      \n",
       "21669                    3           8       1  \n",
       "54843                    3           8       1  \n",
       "77292                    6           8       1  \n",
       "114021                   5           8       1  \n",
       "146670                   6           8       1  \n",
       "\n",
       "[5 rows x 24 columns]"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_df_with_new_features.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Calculating ROC AUC without scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.13543115879019754\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['start_month', 'start_hour']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9536741156181616\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['start_hour', 'morning', 'n_unique_visits']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.944490638088199\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['morning', 'n_unique_visits']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.952393676120362\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['start_hour', 'n_unique_visits']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9163665589418817\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['n_unique_visits']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.944490638088199\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['morning','n_unique_visits']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.13422212319271143\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['start_month','start_hour', 'morning', 'n_unique_visits']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9472937317942698\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['morning']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9561191331053239\n"
     ]
    }
   ],
   "source": [
    "tmp = full_df_with_new_features[['start_hour']].values\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute the metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Calculating ROC AUC with scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9553219908313165\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_df_with_new_features[['start_month', 'start_hour', 'morning', 'n_unique_visits']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9536261104589655\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_df_with_new_features[['start_hour', 'morning', 'n_unique_visits']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9440069634653565\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_df_with_new_features[['morning', 'n_unique_visits']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9517895357206692\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_df_with_new_features[['start_hour',  'n_unique_visits']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9572502735388317\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_df_with_new_features[['start_hour',  'morning']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.947403479438092\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_df_with_new_features[['morning']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9565884665642557\n"
     ]
    }
   ],
   "source": [
    "# Add the new standardized feature to the sparse matrix\n",
    "tmp = StandardScaler().fit_transform(full_df_with_new_features[['start_hour']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], tmp[:idx_split,:]]))\n",
    "\n",
    "# Compute metric on the validation set\n",
    "print(get_auc_lr_valid(X_train, y_train))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## regularization and tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 0 ns\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9591516099541565\n"
     ]
    }
   ],
   "source": [
    "# Compose the training set\n",
    "%time\n",
    "tmp_scaled = StandardScaler().fit_transform(full_df_with_new_features[['start_month', \n",
    "                                                           'start_hour', \n",
    "                                                           'morning']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], \n",
    "                             tmp_scaled[:idx_split,:]]))\n",
    "\n",
    "# Capture the quality with default parameters\n",
    "score_C_1 = get_auc_lr_valid(X_train, y_train)\n",
    "print(score_C_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the C value: 0.001 and its score: 0.8230\n",
      "the C value: 0.003 and its score: 0.8965\n",
      "the C value: 0.008 and its score: 0.9390\n",
      "the C value: 0.022 and its score: 0.9564\n",
      "the C value: 0.060 and its score: 0.9607\n",
      "the C value: 0.167 and its score: 0.9612\n",
      "the C value: 0.464 and its score: 0.9603\n",
      "the C value: 1.292 and its score: 0.9587\n",
      "the C value: 3.594 and its score: 0.9558\n",
      "the C value: 10.000 and its score: 0.9513\n"
     ]
    }
   ],
   "source": [
    "Cs = np.logspace(-3, 1, 10)\n",
    "score_dict = dict()\n",
    "for c in Cs:\n",
    "    each_c_score = get_auc_lr_valid(X_train, y_train,C=c)\n",
    "    score_dict[c] = each_c_score\n",
    "    print(\"the C value: {0:.3f} and its score: {1:.4f}\".format(c,each_c_score))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(Cs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys([0.001, 0.0027825594022071257, 0.007742636826811269, 0.021544346900318832, 0.05994842503189409, 0.1668100537200059, 0.46415888336127775, 1.2915496650148828, 3.593813663804626, 10.0])"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score_dict.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 581,
       "width": 1181
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,ax = plt.subplots(figsize=(20,10))\n",
    "plt.plot( list(score_dict.keys()), list(score_dict.values()))\n",
    "plt.xlim(0.15,0.18)\n",
    "plt.ylim(0.95,0.957)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\python3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\python3\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    }
   ],
   "source": [
    "# Prepare the training and test data\n",
    "tmp_scaled = StandardScaler().fit_transform(full_df_with_new_features[['start_month', 'start_hour', \n",
    "                                                           'morning']])\n",
    "X_train = csr_matrix(hstack([full_sites_sparse[:idx_split,:], \n",
    "                             tmp_scaled[:idx_split,:]]))\n",
    "X_test = csr_matrix(hstack([full_sites_sparse[idx_split:,:], \n",
    "                            tmp_scaled[idx_split:,:]]))\n",
    "\n",
    "# Train the model on the whole training data set using optimal regularization parameter\n",
    "lr = LogisticRegression(C=0.1668100537200059, random_state=17, solver='liblinear').fit(X_train, y_train)\n",
    "\n",
    "# Make a prediction for the test set\n",
    "y_test = lr.predict_proba(X_test)[:, 1]\n",
    "\n",
    "# Write it to the submission file\n",
    "write_to_submission_file(y_test, 'baseline_2.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 2\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "def impute_nan_with_median(table):\n",
    "    for col in table.columns:\n",
    "        table[col]= table[col].fillna(table[col].median())\n",
    "    return table "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SeriousDlqin2yrs</th>\n",
       "      <th>age</th>\n",
       "      <th>NumberOfTime30-59DaysPastDueNotWorse</th>\n",
       "      <th>DebtRatio</th>\n",
       "      <th>NumberOfTimes90DaysLate</th>\n",
       "      <th>NumberOfTime60-89DaysPastDueNotWorse</th>\n",
       "      <th>MonthlyIncome</th>\n",
       "      <th>NumberOfDependents</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>64</td>\n",
       "      <td>0</td>\n",
       "      <td>0.249908</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8158.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>58</td>\n",
       "      <td>0</td>\n",
       "      <td>3870.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>41</td>\n",
       "      <td>0</td>\n",
       "      <td>0.456127</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6666.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>43</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000190</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>10500.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>49</td>\n",
       "      <td>0</td>\n",
       "      <td>0.271820</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>400.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SeriousDlqin2yrs  age  NumberOfTime30-59DaysPastDueNotWorse    DebtRatio  \\\n",
       "0                 0   64                                     0     0.249908   \n",
       "1                 0   58                                     0  3870.000000   \n",
       "2                 0   41                                     0     0.456127   \n",
       "3                 0   43                                     0     0.000190   \n",
       "4                 1   49                                     0     0.271820   \n",
       "\n",
       "   NumberOfTimes90DaysLate  NumberOfTime60-89DaysPastDueNotWorse  \\\n",
       "0                        0                                     0   \n",
       "1                        0                                     0   \n",
       "2                        0                                     0   \n",
       "3                        0                                     0   \n",
       "4                        0                                     0   \n",
       "\n",
       "   MonthlyIncome  NumberOfDependents  \n",
       "0         8158.0                 0.0  \n",
       "1            NaN                 0.0  \n",
       "2         6666.0                 0.0  \n",
       "3        10500.0                 2.0  \n",
       "4          400.0                 0.0  "
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('credit_scoring_sample.csv', sep=\",\",header='infer')\n",
    "data['DebtRatio'] = data['DebtRatio'].replace(',','.', regex=True).astype(float)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ax = data['SeriousDlqin2yrs'].hist(orientation='horizontal', color='red')\n",
    "ax.set_xlabel(\"number_of_observations\")\n",
    "ax.set_ylabel(\"unique_value\")\n",
    "ax.set_title(\"Target distribution\")\n",
    "\n",
    "print('Distribution of target:')\n",
    "data['SeriousDlqin2yrs'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "independent_columns_names = data.columns.values\n",
    "independent_columns_names = [x for x in data if x != 'SeriousDlqin2yrs']\n",
    "independent_columns_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "table = impute_nan_with_median(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = table[independent_columns_names]\n",
    "y = table['SeriousDlqin2yrs']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Question 8. Make an interval estimate based on the bootstrap (2000 samples) of the average income (MonthlyIncome) of customers who had overdue loan payments, and of those who paid in time, make 80% confidence interval. Use target value (SeriousDlqin2yrs) to split data. Find the difference between the lower limit of the derived interval for those who paid in time and the upper limit for those who are overdue. So, you are asked to build 80% intervals for the income of \"good\" customers [good_income_lower,good_income_upper] and for \"bad\" - [bad_income_lower,bad_income_upper] and find the difference good_income_lower−bad_income_upper.\n",
    "\n",
    "Use the example from the article. Set np.random.seed(17). Round your answer to the closest integer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "invalid syntax (<ipython-input-10-7bc9bbd383cb>, line 16)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;36m  File \u001b[1;32m\"<ipython-input-10-7bc9bbd383cb>\"\u001b[1;36m, line \u001b[1;32m16\u001b[0m\n\u001b[1;33m    np.random.seed(17)\u001b[0m\n\u001b[1;37m     ^\u001b[0m\n\u001b[1;31mSyntaxError\u001b[0m\u001b[1;31m:\u001b[0m invalid syntax\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import np.random\n",
    "def get_bootstrap_samples(data, n_samples):\n",
    "    \"\"\"Generate bootstrap samples using the bootstrap method.\"\"\"\n",
    "    indices = np.random.randint(0, len(data), (n_samples, len(data)))\n",
    "    samples = data[indices]\n",
    "    return samples\n",
    "def stat_intervals(stat, alpha):\n",
    "    \"\"\"Produce an interval estimate.\"\"\"\n",
    "    boundaries = np.percentile(stat, [100 * alpha / 2., 100 * (1 - alpha / 2.)])\n",
    "    return boundaries\n",
    "# Save the data about the loyal and former customers to split the dataset\n",
    "good_income = data[data['SeriousDlqin2yrs']== 0]['MonthlyIncome'].values\n",
    "bad_income = data[[['SeriousDlqin2yrs']== 1]['MonthlyIncome'].values\n",
    "# Set the seed for reproducibility of the results\n",
    "np.random.seed(17)\n",
    "# Generate the samples using bootstrapping and calculate the mean for each of them\n",
    "loyal_mean_scores = [np.mean(sample) \n",
    "                       for sample in get_bootstrap_samples(good_income, 2000)]\n",
    "churn_mean_scores = [np.mean(sample) \n",
    "                       for sample in get_bootstrap_samples(bad_income, 2000)]\n",
    "# Print the resulting interval estimates\n",
    "print(\"Service calls from loyal: mean interval\", \n",
    "      stat_intervals(loyal_mean_scores, 0.05))\n",
    "print(\"Service calls from churn: mean interval\", \n",
    "      stat_intervals(churn_mean_scores, 0.05))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}