module3and4.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "f21b84bc",
   "metadata": {},
   "source": [
    "# Segmenting a Database Retrospectively"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cac78b2e",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Computing RFM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "2accfb33",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import needed packages\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import patsy\n",
    "import squarify\n",
    "import statsmodels.api as sm\n",
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ab13f889",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pandasql import sqldf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "30d31a96",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set notebook environment\n",
    "pysqldf = lambda q: sqldf(q, globals())\n",
    "%precision %.2f\n",
    "pd.options.display.float_format = \"{:,.2f}\".format\n",
    "plt.rcParams[\"figure.figsize\"] = 10, 8"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "10e3d959",
   "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>customer_id</th>\n",
       "      <th>purchase_amount</th>\n",
       "      <th>date_of_purchase</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4510</th>\n",
       "      <td>8060</td>\n",
       "      <td>30.00</td>\n",
       "      <td>2014-12-24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17761</th>\n",
       "      <td>109180</td>\n",
       "      <td>50.00</td>\n",
       "      <td>2009-11-25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39110</th>\n",
       "      <td>9830</td>\n",
       "      <td>30.00</td>\n",
       "      <td>2007-06-12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37183</th>\n",
       "      <td>56400</td>\n",
       "      <td>60.00</td>\n",
       "      <td>2009-09-30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33705</th>\n",
       "      <td>41290</td>\n",
       "      <td>60.00</td>\n",
       "      <td>2007-08-21</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       customer_id  purchase_amount date_of_purchase\n",
       "4510          8060            30.00       2014-12-24\n",
       "17761       109180            50.00       2009-11-25\n",
       "39110         9830            30.00       2007-06-12\n",
       "37183        56400            60.00       2009-09-30\n",
       "33705        41290            60.00       2007-08-21"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load text file into a local variable\n",
    "columns = [\"customer_id\", \"purchase_amount\", \"date_of_purchase\"]\n",
    "df = pd.read_csv(\"purchases.txt\", header=None, sep=\"\\t\", names=columns)\n",
    "df.sample(n=5, random_state=57)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "f6e33568",
   "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>customer_id</th>\n",
       "      <th>purchase_amount</th>\n",
       "      <th>date_of_purchase</th>\n",
       "      <th>year_of_purchase</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>760</td>\n",
       "      <td>25.00</td>\n",
       "      <td>2009-11-06</td>\n",
       "      <td>2009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>860</td>\n",
       "      <td>50.00</td>\n",
       "      <td>2012-09-28</td>\n",
       "      <td>2012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1200</td>\n",
       "      <td>100.00</td>\n",
       "      <td>2005-10-25</td>\n",
       "      <td>2005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1420</td>\n",
       "      <td>50.00</td>\n",
       "      <td>2009-07-09</td>\n",
       "      <td>2009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1940</td>\n",
       "      <td>70.00</td>\n",
       "      <td>2013-01-25</td>\n",
       "      <td>2013</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id  purchase_amount date_of_purchase  year_of_purchase\n",
       "0          760            25.00       2009-11-06              2009\n",
       "1          860            50.00       2012-09-28              2012\n",
       "2         1200           100.00       2005-10-25              2005\n",
       "3         1420            50.00       2009-07-09              2009\n",
       "4         1940            70.00       2013-01-25              2013"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# interpret the last column as datetime\n",
    "df[\"date_of_purchase\"] = pd.to_datetime(df[\"date_of_purchase\"], format=\"%Y-%m-%d\")\n",
    "# Extract year of purchase and save it as a column\n",
    "df[\"year_of_purchase\"] = df[\"date_of_purchase\"].dt.year\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "41b8ef4e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add a day_since column showing the difference between last purchase and a basedate\n",
    "basedate = pd.Timestamp(\"2016-01-01\")\n",
    "df[\"days_since\"] = (basedate - df[\"date_of_purchase\"]).dt.days"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b2375709",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 51243 entries, 0 to 51242\n",
      "Data columns (total 5 columns):\n",
      " #   Column            Non-Null Count  Dtype         \n",
      "---  ------            --------------  -----         \n",
      " 0   customer_id       51243 non-null  int64         \n",
      " 1   purchase_amount   51243 non-null  float64       \n",
      " 2   date_of_purchase  51243 non-null  datetime64[ns]\n",
      " 3   year_of_purchase  51243 non-null  int64         \n",
      " 4   days_since        51243 non-null  int64         \n",
      "dtypes: datetime64[ns](1), float64(1), int64(3)\n",
      "memory usage: 2.0 MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d78abf7c",
   "metadata": {},
   "source": [
    "## 2014 - Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "c9cea851",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compute RFM variables as of a year ago using SQL\n",
    "q = \"\"\"\n",
    "        SELECT customer_id,\n",
    "        MIN(days_since) - 365 AS 'recency',\n",
    "        MAX(days_since) - 365 AS 'first_purchase',\n",
    "        COUNT(*) AS 'frequency',\n",
    "        AVG(purchase_amount) AS 'amount'\n",
    "        FROM df\n",
    "        WHERE days_since > 365\n",
    "        GROUP BY 1\"\"\"\n",
    "\n",
    "customers_2014 = sqldf(q)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "0c8155cb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 16905 entries, 0 to 16904\n",
      "Data columns (total 5 columns):\n",
      " #   Column          Non-Null Count  Dtype  \n",
      "---  ------          --------------  -----  \n",
      " 0   customer_id     16905 non-null  int64  \n",
      " 1   recency         16905 non-null  int64  \n",
      " 2   first_purchase  16905 non-null  int64  \n",
      " 3   frequency       16905 non-null  int64  \n",
      " 4   amount          16905 non-null  float64\n",
      "dtypes: float64(1), int64(4)\n",
      "memory usage: 660.5 KB\n"
     ]
    }
   ],
   "source": [
    "customers_2014.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "82e75239",
   "metadata": {},
   "outputs": [],
   "source": [
    "customers_2014.loc[customers_2014[\"recency\"] > 365 * 3, \"segment\"] = \"inactive\"\n",
    "customers_2014[\"segment\"] = customers_2014[\"segment\"].fillna(\"NA\")\n",
    "customers_2014.loc[\n",
    "    (customers_2014[\"recency\"] <= 365 * 3) & (customers_2014[\"recency\"] > 356 * 2),\n",
    "    \"segment\",\n",
    "] = \"cold\"\n",
    "customers_2014.loc[\n",
    "    (customers_2014[\"recency\"] <= 365 * 2) & (customers_2014[\"recency\"] > 365 * 1),\n",
    "    \"segment\",\n",
    "] = \"warm\"\n",
    "customers_2014.loc[customers_2014[\"recency\"] <= 365, \"segment\"] = \"active\"\n",
    "customers_2014.loc[\n",
    "    (customers_2014[\"segment\"] == \"warm\")\n",
    "    & (customers_2014[\"first_purchase\"] <= 365 * 2),\n",
    "    \"segment\",\n",
    "] = \"new warm\"\n",
    "customers_2014.loc[\n",
    "    (customers_2014[\"segment\"] == \"warm\") & (customers_2014[\"amount\"] < 100), \"segment\"\n",
    "] = \"warm low value\"\n",
    "\n",
    "customers_2014.loc[\n",
    "    (customers_2014[\"segment\"] == \"warm\") & (customers_2014[\"amount\"] >= 100), \"segment\"\n",
    "] = \"warm high value\"\n",
    "customers_2014.loc[\n",
    "    (customers_2014[\"segment\"] == \"active\") & (customers_2014[\"first_purchase\"] <= 365),\n",
    "    \"segment\",\n",
    "] = \"new active\"\n",
    "\n",
    "customers_2014.loc[\n",
    "    (customers_2014[\"segment\"] == \"active\") & (customers_2014[\"amount\"] < 100),\n",
    "    \"segment\",\n",
    "] = \"active low value\"\n",
    "\n",
    "customers_2014.loc[\n",
    "    (customers_2014[\"segment\"] == \"active\") & (customers_2014[\"amount\"] >= 100),\n",
    "    \"segment\",\n",
    "] = \"active high value\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "ebb12884",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "inactive             7512\n",
       "active low value     3011\n",
       "cold                 2153\n",
       "new active           1437\n",
       "new warm             1250\n",
       "warm low value        956\n",
       "active high value     475\n",
       "warm high value       111\n",
       "Name: segment, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_2014[\"segment\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "3b9a1dae",
   "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>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1,122.59</td>\n",
       "      <td>1,788.41</td>\n",
       "      <td>2.66</td>\n",
       "      <td>55.53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>964.81</td>\n",
       "      <td>1,024.34</td>\n",
       "      <td>2.72</td>\n",
       "      <td>144.49</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1,868.00</td>\n",
       "      <td>2,695.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>50.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>3,649.00</td>\n",
       "      <td>3,651.00</td>\n",
       "      <td>40.00</td>\n",
       "      <td>4,500.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      recency  first_purchase  frequency   amount\n",
       "mean 1,122.59        1,788.41       2.66    55.53\n",
       "std    964.81        1,024.34       2.72   144.49\n",
       "75%  1,868.00        2,695.00       3.00    50.00\n",
       "max  3,649.00        3,651.00      40.00 4,500.00"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_2014.iloc[:, 1:].describe().loc[[\"mean\", \"std\", \"75%\", \"max\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "8e86c38d",
   "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>customer_id</th>\n",
       "      <th>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>amount</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>3464</td>\n",
       "      <td>3464</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>80</td>\n",
       "      <td>302</td>\n",
       "      <td>3386</td>\n",
       "      <td>6</td>\n",
       "      <td>70.00</td>\n",
       "      <td>active low value</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>90</td>\n",
       "      <td>393</td>\n",
       "      <td>3418</td>\n",
       "      <td>10</td>\n",
       "      <td>115.80</td>\n",
       "      <td>warm high value</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>120</td>\n",
       "      <td>1036</td>\n",
       "      <td>1036</td>\n",
       "      <td>1</td>\n",
       "      <td>20.00</td>\n",
       "      <td>cold</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>130</td>\n",
       "      <td>2605</td>\n",
       "      <td>3345</td>\n",
       "      <td>2</td>\n",
       "      <td>50.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id  recency  first_purchase  frequency  amount           segment\n",
       "0           10     3464            3464          1   30.00          inactive\n",
       "1           80      302            3386          6   70.00  active low value\n",
       "2           90      393            3418         10  115.80   warm high value\n",
       "3          120     1036            1036          1   20.00              cold\n",
       "4          130     2605            3345          2   50.00          inactive"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_2014.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c1c342ba",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Transform segment column datatype from object to category\n",
    "customers_2014[\"segment\"] = customers_2014[\"segment\"].astype(\"category\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7e2e79d9",
   "metadata": {
    "tags": []
   },
   "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>customer_id</th>\n",
       "      <th>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>amount</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>3464</td>\n",
       "      <td>3464</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6872</th>\n",
       "      <td>108170</td>\n",
       "      <td>2261</td>\n",
       "      <td>2261</td>\n",
       "      <td>1</td>\n",
       "      <td>20.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6873</th>\n",
       "      <td>108180</td>\n",
       "      <td>1693</td>\n",
       "      <td>2261</td>\n",
       "      <td>2</td>\n",
       "      <td>20.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6874</th>\n",
       "      <td>108190</td>\n",
       "      <td>2261</td>\n",
       "      <td>2261</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6875</th>\n",
       "      <td>108200</td>\n",
       "      <td>2107</td>\n",
       "      <td>2261</td>\n",
       "      <td>3</td>\n",
       "      <td>36.67</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15939</th>\n",
       "      <td>221470</td>\n",
       "      <td>236</td>\n",
       "      <td>236</td>\n",
       "      <td>1</td>\n",
       "      <td>15.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15938</th>\n",
       "      <td>221460</td>\n",
       "      <td>236</td>\n",
       "      <td>236</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15937</th>\n",
       "      <td>221450</td>\n",
       "      <td>236</td>\n",
       "      <td>236</td>\n",
       "      <td>1</td>\n",
       "      <td>45.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15935</th>\n",
       "      <td>221430</td>\n",
       "      <td>236</td>\n",
       "      <td>236</td>\n",
       "      <td>1</td>\n",
       "      <td>45.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16904</th>\n",
       "      <td>245840</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "      <td>1</td>\n",
       "      <td>50.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>16905 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       customer_id  recency  first_purchase  frequency  amount     segment\n",
       "0               10     3464            3464          1   30.00    inactive\n",
       "6872        108170     2261            2261          1   20.00    inactive\n",
       "6873        108180     1693            2261          2   20.00    inactive\n",
       "6874        108190     2261            2261          1   30.00    inactive\n",
       "6875        108200     2107            2261          3   36.67    inactive\n",
       "...            ...      ...             ...        ...     ...         ...\n",
       "15939       221470      236             236          1   15.00  new active\n",
       "15938       221460      236             236          1   30.00  new active\n",
       "15937       221450      236             236          1   45.00  new active\n",
       "15935       221430      236             236          1   45.00  new active\n",
       "16904       245840       48              48          1   50.00  new active\n",
       "\n",
       "[16905 rows x 6 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Re-order segments in a better readable way\n",
    "sorter = [\n",
    "    \"inactive\",\n",
    "    \"cold\",\n",
    "    \"warm high value\",\n",
    "    \"warm low value\",\n",
    "    \"new warm\",\n",
    "    \"active high value\",\n",
    "    \"active low value\",\n",
    "    \"new active\",\n",
    "]\n",
    "customers_2014.segment.cat.set_categories(sorter, inplace=True)\n",
    "customers_2014.sort_values([\"segment\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "aac29a82",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "c2014 = pd.DataFrame(customers_2014[\"segment\"].value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "90887466",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.0, 100.0, 0.0, 100.0)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "norm = mpl.colors.Normalize(vmin=min(c2014[\"segment\"]), vmax=max(c2014[\"segment\"]))\n",
    "colors = [mpl.cm.Blues(norm(value)) for value in c2014[\"segment\"]]\n",
    "squarify.plot(sizes=c2014[\"segment\"], label=c2014.index, color=colors, alpha=0.6)\n",
    "plt.axis(\"off\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8fd4e1b1",
   "metadata": {},
   "source": [
    "## 2015 - Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "923ec778",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compute revenues generated by customers in 2015 using SQL\n",
    "q = \"\"\"\n",
    "        SELECT customer_id, \n",
    "        SUM(purchase_amount) AS 'revenue_2015'\n",
    "        FROM df\n",
    "        WHERE year_of_purchase = 2015\n",
    "        GROUP BY 1\n",
    "    \"\"\"\n",
    "revenue_2015 = sqldf(q)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "565adc50",
   "metadata": {},
   "outputs": [],
   "source": [
    "q = \"\"\"\n",
    "        SELECT customer_id,\n",
    "        MIN(days_since) AS 'recency',\n",
    "        MAX(days_since) AS 'first_purchase',\n",
    "        COUNT(*) AS 'frequency',\n",
    "        AVG(purchase_amount) AS 'amount'\n",
    "        FROM df GROUP BY 1\"\"\"\n",
    "customers_2015 = sqldf(q)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "e8ed67bf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5398"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "revenue_2015.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "44b0eb94",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18417"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_2015.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "10e604f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "customers_2015.loc[customers_2015[\"recency\"] > 365 * 3, \"segment\"] = \"inactive\"\n",
    "customers_2015[\"segment\"] = customers_2015[\"segment\"].fillna(\"NA\")\n",
    "customers_2015.loc[\n",
    "    (customers_2015[\"recency\"] <= 365 * 3) & (customers_2015[\"recency\"] > 356 * 2),\n",
    "    \"segment\",\n",
    "] = \"cold\"\n",
    "customers_2015.loc[\n",
    "    (customers_2015[\"recency\"] <= 365 * 2) & (customers_2015[\"recency\"] > 365 * 1),\n",
    "    \"segment\",\n",
    "] = \"warm\"\n",
    "customers_2015.loc[customers_2015[\"recency\"] <= 365, \"segment\"] = \"active\"\n",
    "customers_2015.loc[\n",
    "    (customers_2015[\"segment\"] == \"warm\")\n",
    "    & (customers_2015[\"first_purchase\"] <= 365 * 2),\n",
    "    \"segment\",\n",
    "] = \"new warm\"\n",
    "customers_2015.loc[\n",
    "    (customers_2015[\"segment\"] == \"warm\") & (customers_2015[\"amount\"] < 100), \"segment\"\n",
    "] = \"warm low value\"\n",
    "customers_2015.loc[\n",
    "    (customers_2015[\"segment\"] == \"warm\") & (customers_2015[\"amount\"] >= 100), \"segment\"\n",
    "] = \"warm high value\"\n",
    "customers_2015.loc[\n",
    "    (customers_2015[\"segment\"] == \"active\") & (customers_2015[\"first_purchase\"] <= 365),\n",
    "    \"segment\",\n",
    "] = \"new active\"\n",
    "customers_2015.loc[\n",
    "    (customers_2015[\"segment\"] == \"active\") & (customers_2015[\"amount\"] < 100),\n",
    "    \"segment\",\n",
    "] = \"active low value\"\n",
    "customers_2015.loc[\n",
    "    (customers_2015[\"segment\"] == \"active\") & (customers_2015[\"amount\"] >= 100),\n",
    "    \"segment\",\n",
    "] = \"active high value\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "18d77675",
   "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>customer_id</th>\n",
       "      <th>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>amount</th>\n",
       "      <th>segment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>3829</td>\n",
       "      <td>3829</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7894</th>\n",
       "      <td>119690</td>\n",
       "      <td>1500</td>\n",
       "      <td>2443</td>\n",
       "      <td>4</td>\n",
       "      <td>30.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7895</th>\n",
       "      <td>119710</td>\n",
       "      <td>2249</td>\n",
       "      <td>2443</td>\n",
       "      <td>3</td>\n",
       "      <td>26.67</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7896</th>\n",
       "      <td>119720</td>\n",
       "      <td>2156</td>\n",
       "      <td>2443</td>\n",
       "      <td>2</td>\n",
       "      <td>40.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7897</th>\n",
       "      <td>119730</td>\n",
       "      <td>2443</td>\n",
       "      <td>2443</td>\n",
       "      <td>1</td>\n",
       "      <td>50.00</td>\n",
       "      <td>inactive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17404</th>\n",
       "      <td>252380</td>\n",
       "      <td>73</td>\n",
       "      <td>73</td>\n",
       "      <td>1</td>\n",
       "      <td>10.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17403</th>\n",
       "      <td>252370</td>\n",
       "      <td>73</td>\n",
       "      <td>73</td>\n",
       "      <td>1</td>\n",
       "      <td>45.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17402</th>\n",
       "      <td>252360</td>\n",
       "      <td>73</td>\n",
       "      <td>73</td>\n",
       "      <td>1</td>\n",
       "      <td>20.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17411</th>\n",
       "      <td>252450</td>\n",
       "      <td>72</td>\n",
       "      <td>72</td>\n",
       "      <td>1</td>\n",
       "      <td>45.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18416</th>\n",
       "      <td>264200</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>500.00</td>\n",
       "      <td>new active</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>18417 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       customer_id  recency  first_purchase  frequency  amount     segment\n",
       "0               10     3829            3829          1   30.00    inactive\n",
       "7894        119690     1500            2443          4   30.00    inactive\n",
       "7895        119710     2249            2443          3   26.67    inactive\n",
       "7896        119720     2156            2443          2   40.00    inactive\n",
       "7897        119730     2443            2443          1   50.00    inactive\n",
       "...            ...      ...             ...        ...     ...         ...\n",
       "17404       252380       73              73          1   10.00  new active\n",
       "17403       252370       73              73          1   45.00  new active\n",
       "17402       252360       73              73          1   20.00  new active\n",
       "17411       252450       72              72          1   45.00  new active\n",
       "18416       264200        1               1          1  500.00  new active\n",
       "\n",
       "[18417 rows x 6 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Transform segment column datatype from object to category\n",
    "customers_2015[\"segment\"] = customers_2015[\"segment\"].astype(\"category\")\n",
    "# Re-order segments in a better readable way\n",
    "sorter = [\n",
    "    \"inactive\",\n",
    "    \"cold\",\n",
    "    \"warm high value\",\n",
    "    \"warm low value\",\n",
    "    \"new warm\",\n",
    "    \"active high value\",\n",
    "    \"active low value\",\n",
    "    \"new active\",\n",
    "]\n",
    "customers_2015.segment.cat.set_categories(sorter, inplace=True)\n",
    "customers_2015.sort_values([\"segment\"], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "1dd9b27c",
   "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>customer_id</th>\n",
       "      <th>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>amount</th>\n",
       "      <th>revenue_2015</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>18,417.00</td>\n",
       "      <td>18,417.00</td>\n",
       "      <td>18,417.00</td>\n",
       "      <td>18,417.00</td>\n",
       "      <td>18,417.00</td>\n",
       "      <td>5,398.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>137,573.51</td>\n",
       "      <td>1,253.04</td>\n",
       "      <td>1,984.01</td>\n",
       "      <td>2.78</td>\n",
       "      <td>57.79</td>\n",
       "      <td>88.62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>69,504.60</td>\n",
       "      <td>1,081.44</td>\n",
       "      <td>1,133.41</td>\n",
       "      <td>2.94</td>\n",
       "      <td>154.36</td>\n",
       "      <td>224.36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>10.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>5.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>81,990.00</td>\n",
       "      <td>244.00</td>\n",
       "      <td>988.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>21.67</td>\n",
       "      <td>30.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>136,430.00</td>\n",
       "      <td>1,070.00</td>\n",
       "      <td>2,087.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>30.00</td>\n",
       "      <td>50.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>195,100.00</td>\n",
       "      <td>2,130.00</td>\n",
       "      <td>2,992.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>50.00</td>\n",
       "      <td>85.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>264,200.00</td>\n",
       "      <td>4,014.00</td>\n",
       "      <td>4,016.00</td>\n",
       "      <td>45.00</td>\n",
       "      <td>4,500.00</td>\n",
       "      <td>4,500.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       customer_id   recency  first_purchase  frequency    amount  \\\n",
       "count    18,417.00 18,417.00       18,417.00  18,417.00 18,417.00   \n",
       "mean    137,573.51  1,253.04        1,984.01       2.78     57.79   \n",
       "std      69,504.60  1,081.44        1,133.41       2.94    154.36   \n",
       "min          10.00      1.00            1.00       1.00      5.00   \n",
       "25%      81,990.00    244.00          988.00       1.00     21.67   \n",
       "50%     136,430.00  1,070.00        2,087.00       2.00     30.00   \n",
       "75%     195,100.00  2,130.00        2,992.00       3.00     50.00   \n",
       "max     264,200.00  4,014.00        4,016.00      45.00  4,500.00   \n",
       "\n",
       "       revenue_2015  \n",
       "count      5,398.00  \n",
       "mean          88.62  \n",
       "std          224.36  \n",
       "min            5.00  \n",
       "25%           30.00  \n",
       "50%           50.00  \n",
       "75%           85.00  \n",
       "max        4,500.00  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merge 2015 customers and 2015 revenue, while making sure that customers\n",
    "# without purchase in 2015 also appear in the new df\n",
    "actual = customers_2015.merge(revenue_2015, how=\"left\", on=\"customer_id\")\n",
    "actual.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "42409e46",
   "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>customer_id</th>\n",
       "      <th>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>amount</th>\n",
       "      <th>segment</th>\n",
       "      <th>revenue_2015</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>3829</td>\n",
       "      <td>3829</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "      <td>inactive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>80</td>\n",
       "      <td>343</td>\n",
       "      <td>3751</td>\n",
       "      <td>7</td>\n",
       "      <td>71.43</td>\n",
       "      <td>active low value</td>\n",
       "      <td>80.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>90</td>\n",
       "      <td>758</td>\n",
       "      <td>3783</td>\n",
       "      <td>10</td>\n",
       "      <td>115.80</td>\n",
       "      <td>cold</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>120</td>\n",
       "      <td>1401</td>\n",
       "      <td>1401</td>\n",
       "      <td>1</td>\n",
       "      <td>20.00</td>\n",
       "      <td>inactive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>130</td>\n",
       "      <td>2970</td>\n",
       "      <td>3710</td>\n",
       "      <td>2</td>\n",
       "      <td>50.00</td>\n",
       "      <td>inactive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id  recency  first_purchase  frequency  amount           segment  \\\n",
       "0           10     3829            3829          1   30.00          inactive   \n",
       "1           80      343            3751          7   71.43  active low value   \n",
       "2           90      758            3783         10  115.80              cold   \n",
       "3          120     1401            1401          1   20.00          inactive   \n",
       "4          130     2970            3710          2   50.00          inactive   \n",
       "\n",
       "   revenue_2015  \n",
       "0           NaN  \n",
       "1         80.00  \n",
       "2           NaN  \n",
       "3           NaN  \n",
       "4           NaN  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actual.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "8d9e96dd",
   "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>customer_id</th>\n",
       "      <th>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>amount</th>\n",
       "      <th>segment</th>\n",
       "      <th>revenue_2015</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>3829</td>\n",
       "      <td>3829</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "      <td>inactive</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>80</td>\n",
       "      <td>343</td>\n",
       "      <td>3751</td>\n",
       "      <td>7</td>\n",
       "      <td>71.43</td>\n",
       "      <td>active low value</td>\n",
       "      <td>80.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>90</td>\n",
       "      <td>758</td>\n",
       "      <td>3783</td>\n",
       "      <td>10</td>\n",
       "      <td>115.80</td>\n",
       "      <td>cold</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>120</td>\n",
       "      <td>1401</td>\n",
       "      <td>1401</td>\n",
       "      <td>1</td>\n",
       "      <td>20.00</td>\n",
       "      <td>inactive</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>130</td>\n",
       "      <td>2970</td>\n",
       "      <td>3710</td>\n",
       "      <td>2</td>\n",
       "      <td>50.00</td>\n",
       "      <td>inactive</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id  recency  first_purchase  frequency  amount           segment  \\\n",
       "0           10     3829            3829          1   30.00          inactive   \n",
       "1           80      343            3751          7   71.43  active low value   \n",
       "2           90      758            3783         10  115.80              cold   \n",
       "3          120     1401            1401          1   20.00          inactive   \n",
       "4          130     2970            3710          2   50.00          inactive   \n",
       "\n",
       "   revenue_2015  \n",
       "0          0.00  \n",
       "1         80.00  \n",
       "2          0.00  \n",
       "3          0.00  \n",
       "4          0.00  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replace NaNs in the revenus_2015 column with 0s\n",
    "actual[\"revenue_2015\"].fillna(0, inplace=True)\n",
    "actual.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "f7b44c6d",
   "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>revenue_2015</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>segment</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>inactive</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cold</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>warm high value</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>warm low value</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>new warm</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>active high value</th>\n",
       "      <td>323.57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>active low value</th>\n",
       "      <td>52.31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>new active</th>\n",
       "      <td>79.17</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   revenue_2015\n",
       "segment                        \n",
       "inactive                   0.00\n",
       "cold                       0.00\n",
       "warm high value            0.00\n",
       "warm low value             0.00\n",
       "new warm                   0.00\n",
       "active high value        323.57\n",
       "active low value          52.31\n",
       "new active                79.17"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actual[[\"revenue_2015\", \"segment\"]].groupby(\"segment\").mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "532846ab",
   "metadata": {},
   "source": [
    "By definition inactive, warm and cold didn't generate any revenue in the same year we are investigating. That's why the 0.00 in these rows."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2ae8c26",
   "metadata": {},
   "source": [
    "## Computing Predictors and Target Variables\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "51225de6",
   "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>customer_id</th>\n",
       "      <th>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>amount</th>\n",
       "      <th>segment</th>\n",
       "      <th>revenue_2015</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>3464</td>\n",
       "      <td>3464</td>\n",
       "      <td>1</td>\n",
       "      <td>30.00</td>\n",
       "      <td>inactive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>80</td>\n",
       "      <td>302</td>\n",
       "      <td>3386</td>\n",
       "      <td>6</td>\n",
       "      <td>70.00</td>\n",
       "      <td>active low value</td>\n",
       "      <td>80.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>90</td>\n",
       "      <td>393</td>\n",
       "      <td>3418</td>\n",
       "      <td>10</td>\n",
       "      <td>115.80</td>\n",
       "      <td>warm high value</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>120</td>\n",
       "      <td>1036</td>\n",
       "      <td>1036</td>\n",
       "      <td>1</td>\n",
       "      <td>20.00</td>\n",
       "      <td>cold</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>130</td>\n",
       "      <td>2605</td>\n",
       "      <td>3345</td>\n",
       "      <td>2</td>\n",
       "      <td>50.00</td>\n",
       "      <td>inactive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id  recency  first_purchase  frequency  amount           segment  \\\n",
       "0           10     3464            3464          1   30.00          inactive   \n",
       "1           80      302            3386          6   70.00  active low value   \n",
       "2           90      393            3418         10  115.80   warm high value   \n",
       "3          120     1036            1036          1   20.00              cold   \n",
       "4          130     2605            3345          2   50.00          inactive   \n",
       "\n",
       "   revenue_2015  \n",
       "0           NaN  \n",
       "1         80.00  \n",
       "2           NaN  \n",
       "3           NaN  \n",
       "4           NaN  "
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merge 2014 customers and 2015 revenue\n",
    "forward = customers_2014.merge(revenue_2015, how=\"left\", on=\"customer_id\")\n",
    "forward.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "a31fda9b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Replace NaNs in the revenus_2015 column with 0s\n",
    "forward[\"revenue_2015\"].fillna(0, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "cda6ced3",
   "metadata": {},
   "outputs": [],
   "source": [
    "r = forward[[\"revenue_2015\", \"segment\"]].groupby(\"segment\").mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "6b83becf",
   "metadata": {},
   "outputs": [],
   "source": [
    "r.sort_values(by=[\"revenue_2015\"], ascending=False, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "31bd57c8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='segment'>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "r.plot.bar()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "89602ac2",
   "metadata": {},
   "source": [
    "As we can see, customers classified as \"active high value\" in 2014 are the most revenue generating in 2015 and in the second place comes customers from the \"warm high value\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f096303d",
   "metadata": {},
   "source": [
    "## Quiz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "08c6b816",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "inactive             9158\n",
       "active low value     3313\n",
       "cold                 1903\n",
       "new active           1512\n",
       "new warm              938\n",
       "warm low value        901\n",
       "active high value     573\n",
       "warm high value       119\n",
       "Name: segment, dtype: int64"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actual[\"segment\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "e2da0ff5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3313"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How many \"new active low\" (active low value) customers were there in 2015?\n",
    "\n",
    "(actual[\"segment\"] == \"active low value\").sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "c788b1e3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20.63\n"
     ]
    }
   ],
   "source": [
    "# The number of \"new active high\" customers has increased between 2014 and 2015.\n",
    "# What is the rate of that increase?\n",
    "\n",
    "ne = (customers_2015[\"segment\"] == \"active high value\").sum()\n",
    "ol = (customers_2014[\"segment\"] == \"active high value\").sum()\n",
    "\n",
    "print(\"%.2f\" % ((ne - ol) / ol * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "8afb9528",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20.63\n"
     ]
    }
   ],
   "source": [
    "print(\"{:.2f}\".format((ne - ol) / ol * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "2509c966",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20.63157894736842\n"
     ]
    }
   ],
   "source": [
    "print(f\"{(ne - ol)/ol * 100}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "557ae6d2",
   "metadata": {},
   "source": [
    "# Building a Predictive Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "c503d03a",
   "metadata": {},
   "outputs": [],
   "source": [
    "q = \"\"\"\n",
    "        SELECT customer_id, \n",
    "        MIN(days_since) - 365 AS 'recency', \n",
    "        MAX(days_since) - 365 AS 'first_purchase', \n",
    "        COUNT(*) AS 'frequency', \n",
    "        AVG(purchase_amount) AS 'avg_amount', \n",
    "        MAX(purchase_amount) AS 'max_amount' \n",
    "        FROM df WHERE days_since > 365 GROUP BY 1\n",
    "\"\"\"\n",
    "customers_2014 = sqldf(q, globals())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "cba2d715",
   "metadata": {},
   "outputs": [],
   "source": [
    "q = \"\"\"\n",
    "        SELECT customer_id, \n",
    "        SUM(purchase_amount) AS 'revenue_2015' \n",
    "        FROM df WHERE year_of_purchase = 2015 \n",
    "        GROUP BY 1\n",
    "\"\"\"\n",
    "revenue_2015 = sqldf(q, globals())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "6de8fe90",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Merge 2014 customers with 2015 revenue\n",
    "in_sample = customers_2014.merge(revenue_2015, how=\"left\", on=\"customer_id\")\n",
    "in_sample[\"revenue_2015\"].fillna(0, inplace=True)\n",
    "in_sample.loc[in_sample[\"revenue_2015\"] > 0, \"active_2015\"] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "90b312d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 16905 entries, 0 to 16904\n",
      "Data columns (total 8 columns):\n",
      " #   Column          Non-Null Count  Dtype  \n",
      "---  ------          --------------  -----  \n",
      " 0   customer_id     16905 non-null  int64  \n",
      " 1   recency         16905 non-null  int64  \n",
      " 2   first_purchase  16905 non-null  int64  \n",
      " 3   frequency       16905 non-null  int64  \n",
      " 4   avg_amount      16905 non-null  float64\n",
      " 5   max_amount      16905 non-null  float64\n",
      " 6   revenue_2015    16905 non-null  float64\n",
      " 7   active_2015     16905 non-null  int64  \n",
      "dtypes: float64(3), int64(5)\n",
      "memory usage: 1.2 MB\n"
     ]
    }
   ],
   "source": [
    "in_sample[\"active_2015\"].fillna(0, inplace=True)\n",
    "in_sample[\"active_2015\"] = in_sample[\"active_2015\"].astype(\"int\")\n",
    "in_sample.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "776738a4",
   "metadata": {},
   "source": [
    "------"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03234b13",
   "metadata": {},
   "source": [
    "## The Likelihood a Customer will be Active in 2015"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da74d598",
   "metadata": {},
   "source": [
    "Now, we are going to calibrate the first model, which is the probability model. The likelihood that a customer we'd be active in 2015 or not. \n",
    "\n",
    "* the output of the model which I call prob dot model is the output of the Multinom function with formula that states that active 2015 is a function of recency first purchase as frequency, average amount and maximum amount.\n",
    "* And the data is the data we created the calibration data called In sample.\n",
    "\n",
    "These will fit, calibrate the entire model on the data set. And then we can going to extract the coefficients and the standard deviations of these coefficients, and output not only the coefficients, the standard deviations, but the ratio of those as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "0d0a5bce",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.365836\n",
      "         Iterations 8\n"
     ]
    }
   ],
   "source": [
    "# We use the Logit function since the output could be either 0 or 1\n",
    "prob_model = sm.Logit.from_formula(\n",
    "    \"active_2015 ~ recency + first_purchase + frequency + avg_amount + max_amount\",\n",
    "    in_sample,\n",
    ")\n",
    "prob_model_fit = prob_model.fit()\n",
    "coef = prob_model_fit.params\n",
    "std = prob_model_fit.bse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "899b8322",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Intercept        -0.53\n",
      "recency          -0.00\n",
      "first_purchase   -0.00\n",
      "frequency         0.22\n",
      "avg_amount        0.00\n",
      "max_amount       -0.00\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(coef)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "ad203526",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Intercept        0.04\n",
      "recency          0.00\n",
      "first_purchase   0.00\n",
      "frequency        0.01\n",
      "avg_amount       0.00\n",
      "max_amount       0.00\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(std)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "bfd4b466",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Intercept        -12.09\n",
      "recency          -32.75\n",
      "first_purchase    -0.30\n",
      "frequency         14.84\n",
      "avg_amount         1.14\n",
      "max_amount        -0.57\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(coef / std)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c53c5a71",
   "metadata": {},
   "source": [
    "The model converged! 👍\n",
    "\n",
    "The recency parameter, for instance, it's negative. Which makes perfect sense. The larger the recency, meaning the more days have lapsed between the last purchase and today, the less likely you're going to make any other purchase in the future.\n",
    "\n",
    "So, if your last purchase was three four ten years ago, it's extremely unlikely that you'll make any purchase very soon. Meaning that the sign of the primary is negative, the higher the recency, the lower the probability.\n",
    "\n",
    "if you look at frequency, that primary is positive. Meaning that the more purchases you've made in the past, the more likely you'll make additional purchases in the future, which makes perfect sense. Now, these two parameter values are the most interesting, simply because if you look at standard deviations and at the ratio between coefficients and standard deviation, which usually indicate to what extent each parameter value is significant, or not. If it's above two, or below minus two, usually it's a good sign, and as you can see here, recently is huge, minus 32"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9de50685",
   "metadata": {},
   "source": [
    "## Monetary Prediction Model\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "d3ce7bd4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# For the monetary model, we select only those who made a purchase\n",
    "# Only the index are enough\n",
    "z = in_sample[in_sample[\"active_2015\"] == 1].index.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "ac309d93",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1, 17, 29, 30, 31]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "z[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "277d102e",
   "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>customer_id</th>\n",
       "      <th>recency</th>\n",
       "      <th>first_purchase</th>\n",
       "      <th>frequency</th>\n",
       "      <th>avg_amount</th>\n",
       "      <th>max_amount</th>\n",
       "      <th>revenue_2015</th>\n",
       "      <th>active_2015</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>3,886.00</td>\n",
       "      <td>3,886.00</td>\n",
       "      <td>3,886.00</td>\n",
       "      <td>3,886.00</td>\n",
       "      <td>3,886.00</td>\n",
       "      <td>3,886.00</td>\n",
       "      <td>3,886.00</td>\n",
       "      <td>3,886.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>134,906.97</td>\n",
       "      <td>306.35</td>\n",
       "      <td>1,636.16</td>\n",
       "      <td>4.74</td>\n",
       "      <td>67.78</td>\n",
       "      <td>88.33</td>\n",
       "      <td>92.30</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>68,404.30</td>\n",
       "      <td>519.46</td>\n",
       "      <td>1,101.25</td>\n",
       "      <td>3.79</td>\n",
       "      <td>160.06</td>\n",
       "      <td>222.15</td>\n",
       "      <td>217.45</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>80.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>78,590.00</td>\n",
       "      <td>23.00</td>\n",
       "      <td>649.75</td>\n",
       "      <td>2.00</td>\n",
       "      <td>30.00</td>\n",
       "      <td>30.00</td>\n",
       "      <td>30.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>143,550.00</td>\n",
       "      <td>97.00</td>\n",
       "      <td>1,604.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>40.00</td>\n",
       "      <td>50.00</td>\n",
       "      <td>50.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>194,362.50</td>\n",
       "      <td>328.00</td>\n",
       "      <td>2,666.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>60.00</td>\n",
       "      <td>80.00</td>\n",
       "      <td>100.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>236,660.00</td>\n",
       "      <td>3,544.00</td>\n",
       "      <td>3,647.00</td>\n",
       "      <td>40.00</td>\n",
       "      <td>4,500.00</td>\n",
       "      <td>4,500.00</td>\n",
       "      <td>4,500.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       customer_id  recency  first_purchase  frequency  avg_amount  \\\n",
       "count     3,886.00 3,886.00        3,886.00   3,886.00    3,886.00   \n",
       "mean    134,906.97   306.35        1,636.16       4.74       67.78   \n",
       "std      68,404.30   519.46        1,101.25       3.79      160.06   \n",
       "min          80.00     1.00            1.00       1.00        5.00   \n",
       "25%      78,590.00    23.00          649.75       2.00       30.00   \n",
       "50%     143,550.00    97.00        1,604.00       4.00       40.00   \n",
       "75%     194,362.50   328.00        2,666.00       7.00       60.00   \n",
       "max     236,660.00 3,544.00        3,647.00      40.00    4,500.00   \n",
       "\n",
       "       max_amount  revenue_2015  active_2015  \n",
       "count    3,886.00      3,886.00     3,886.00  \n",
       "mean        88.33         92.30         1.00  \n",
       "std        222.15        217.45         0.00  \n",
       "min          5.00          5.00         1.00  \n",
       "25%         30.00         30.00         1.00  \n",
       "50%         50.00         50.00         1.00  \n",
       "75%         80.00        100.00         1.00  \n",
       "max      4,500.00      4,500.00         1.00  "
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "in_sample.loc[z].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "87c39021",
   "metadata": {},
   "source": [
    "We only have active customers in there. In terms of revenue, customers have spent anything between \\\\$5 and \\\\$4500 with that retailer of the year 2015. Now, what you are going to do is calibrate the monetary model. Meaning we begin to predict how much they spend in 2015 based on only two things here, the average amount they spend usually, and the maximum amount they spend. So we have two different predictors."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e94064f",
   "metadata": {},
   "source": [
    "SM, which stands for Linear Model, will fit a linear model to match as closely as we can revenue 2015 based on the predictor's average amount and maximum amount. And the data here is not the entire sample, but only those customers who can be found in the index, Z here. So only those customers who actually spend something."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "a2da547b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>      <td>revenue_2015</td>   <th>  R-squared:         </th> <td>   0.605</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                   <td>OLS</td>       <th>  Adj. R-squared:    </th> <td>   0.605</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>             <td>Least Squares</td>  <th>  F-statistic:       </th> <td>   2979.</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>             <td>Thu, 22 Apr 2021</td> <th>  Prob (F-statistic):</th>  <td>  0.00</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                 <td>15:00:27</td>     <th>  Log-Likelihood:    </th> <td> -24621.</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>      <td>  3886</td>      <th>  AIC:               </th> <td>4.925e+04</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>          <td>  3883</td>      <th>  BIC:               </th> <td>4.927e+04</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>              <td>     2</td>      <th>                     </th>     <td> </td>    \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>      <td>nonrobust</td>    <th>                     </th>     <td> </td>    \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "       <td></td>         <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Intercept</th>  <td>   20.7471</td> <td>    2.381</td> <td>    8.713</td> <td> 0.000</td> <td>   16.079</td> <td>   25.415</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>avg_amount</th> <td>    0.6749</td> <td>    0.033</td> <td>   20.575</td> <td> 0.000</td> <td>    0.611</td> <td>    0.739</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>max_amount</th> <td>    0.2923</td> <td>    0.024</td> <td>   12.367</td> <td> 0.000</td> <td>    0.246</td> <td>    0.339</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td>5580.836</td> <th>  Durbin-Watson:     </th>  <td>   2.007</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th>  <td> 0.000</td>  <th>  Jarque-Bera (JB):  </th> <td>8162692.709</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>           <td> 7.843</td>  <th>  Prob(JB):          </th>  <td>    0.00</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>       <td>226.980</td> <th>  Cond. No.          </th>  <td>    315.</td>  \n",
       "</tr>\n",
       "</table><br/><br/>Notes:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                            OLS Regression Results                            \n",
       "==============================================================================\n",
       "Dep. Variable:           revenue_2015   R-squared:                       0.605\n",
       "Model:                            OLS   Adj. R-squared:                  0.605\n",
       "Method:                 Least Squares   F-statistic:                     2979.\n",
       "Date:                Thu, 22 Apr 2021   Prob (F-statistic):               0.00\n",
       "Time:                        15:00:27   Log-Likelihood:                -24621.\n",
       "No. Observations:                3886   AIC:                         4.925e+04\n",
       "Df Residuals:                    3883   BIC:                         4.927e+04\n",
       "Df Model:                           2                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "==============================================================================\n",
       "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
       "------------------------------------------------------------------------------\n",
       "Intercept     20.7471      2.381      8.713      0.000      16.079      25.415\n",
       "avg_amount     0.6749      0.033     20.575      0.000       0.611       0.739\n",
       "max_amount     0.2923      0.024     12.367      0.000       0.246       0.339\n",
       "==============================================================================\n",
       "Omnibus:                     5580.836   Durbin-Watson:                   2.007\n",
       "Prob(Omnibus):                  0.000   Jarque-Bera (JB):          8162692.709\n",
       "Skew:                           7.843   Prob(JB):                         0.00\n",
       "Kurtosis:                     226.980   Cond. No.                         315.\n",
       "==============================================================================\n",
       "\n",
       "Notes:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "\"\"\""
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calibrate the monetary model (version 1)\n",
    "amount_model = sm.OLS.from_formula(\n",
    "    \"revenue_2015 ~ avg_amount + max_amount\", in_sample.loc[z]\n",
    ")\n",
    "amount_model_fit = amount_model.fit()\n",
    "amount_model_fit.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af3c8c82",
   "metadata": {},
   "source": [
    "Our square value is point 60, which is basically a signal of the fit of the model. But, we have a slight issue here.\n",
    "\n",
    "If we plot on one hand how much has been spent by customers and on the other we take the amount model we've just created through linear regression, and look at the fitted values, we get an ugly chart. And the reason is that most customers have spent small amounts and a few outliers have spent huge amounts, up to \\$4.000."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "67ac6f85",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb8d7f75b80>"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(in_sample.loc[z].revenue_2015, amount_model_fit.fittedvalues)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5d6a616d",
   "metadata": {
    "tags": []
   },
   "source": [
    "So for optimization purposes, instead of creating a model with the amount we're going to create an amount with logaritmic distribution.\n",
    "\n",
    "So instead of predicting revenue 2015 bsed on average amount and maximum amount. We're going to predict the log of revenue 2015 based on the log on average"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "450376d1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>OLS Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>    <td>np.log(revenue_2015)</td> <th>  R-squared:         </th> <td>   0.693</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                     <td>OLS</td>         <th>  Adj. R-squared:    </th> <td>   0.693</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>               <td>Least Squares</td>    <th>  F-statistic:       </th> <td>   4377.</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>               <td>Thu, 22 Apr 2021</td>   <th>  Prob (F-statistic):</th>  <td>  0.00</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                   <td>15:00:28</td>       <th>  Log-Likelihood:    </th> <td> -2644.6</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Observations:</th>        <td>  3886</td>        <th>  AIC:               </th> <td>   5295.</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Residuals:</th>            <td>  3883</td>        <th>  BIC:               </th> <td>   5314.</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Df Model:</th>                <td>     2</td>        <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>        <td>nonrobust</td>      <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "           <td></td>             <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Intercept</th>          <td>    0.3700</td> <td>    0.040</td> <td>    9.242</td> <td> 0.000</td> <td>    0.292</td> <td>    0.448</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>np.log(avg_amount)</th> <td>    0.5488</td> <td>    0.042</td> <td>   13.171</td> <td> 0.000</td> <td>    0.467</td> <td>    0.631</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>np.log(max_amount)</th> <td>    0.3881</td> <td>    0.038</td> <td>   10.224</td> <td> 0.000</td> <td>    0.314</td> <td>    0.463</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "  <th>Omnibus:</th>       <td>501.505</td> <th>  Durbin-Watson:     </th> <td>   1.961</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Prob(Omnibus):</th> <td> 0.000</td>  <th>  Jarque-Bera (JB):  </th> <td>3328.833</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Skew:</th>          <td> 0.421</td>  <th>  Prob(JB):          </th> <td>    0.00</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Kurtosis:</th>      <td> 7.455</td>  <th>  Cond. No.          </th> <td>    42.2</td>\n",
       "</tr>\n",
       "</table><br/><br/>Notes:<br/>[1] Standard Errors assume that the covariance matrix of the errors is correctly specified."
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                             OLS Regression Results                             \n",
       "================================================================================\n",
       "Dep. Variable:     np.log(revenue_2015)   R-squared:                       0.693\n",
       "Model:                              OLS   Adj. R-squared:                  0.693\n",
       "Method:                   Least Squares   F-statistic:                     4377.\n",
       "Date:                  Thu, 22 Apr 2021   Prob (F-statistic):               0.00\n",
       "Time:                          15:00:28   Log-Likelihood:                -2644.6\n",
       "No. Observations:                  3886   AIC:                             5295.\n",
       "Df Residuals:                      3883   BIC:                             5314.\n",
       "Df Model:                             2                                         \n",
       "Covariance Type:              nonrobust                                         \n",
       "======================================================================================\n",
       "                         coef    std err          t      P>|t|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------------\n",
       "Intercept              0.3700      0.040      9.242      0.000       0.292       0.448\n",
       "np.log(avg_amount)     0.5488      0.042     13.171      0.000       0.467       0.631\n",
       "np.log(max_amount)     0.3881      0.038     10.224      0.000       0.314       0.463\n",
       "==============================================================================\n",
       "Omnibus:                      501.505   Durbin-Watson:                   1.961\n",
       "Prob(Omnibus):                  0.000   Jarque-Bera (JB):             3328.833\n",
       "Skew:                           0.421   Prob(JB):                         0.00\n",
       "Kurtosis:                       7.455   Cond. No.                         42.2\n",
       "==============================================================================\n",
       "\n",
       "Notes:\n",
       "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
       "\"\"\""
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Re-calibrate the monetary model, using a log-transform (version 2)\n",
    "amount_model = sm.OLS.from_formula(\n",
    "    \"np.log(revenue_2015) ~ np.log(avg_amount) + np.log(max_amount)\", in_sample.loc[z]\n",
    ")\n",
    "amount_model_fit = amount_model.fit()\n",
    "amount_model_fit.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a410a9ff",
   "metadata": {},
   "source": [
    "The R-squared has improved, meaning we fit the data much better.\n",
    "\n",
    "Now the plot makes much more sense too. Since we put more weight to the smaller values and less weight to the very large values. \n",
    "\n",
    "It looks like we could generate a line through the cloud of points predicting the revenue of 2015."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "4a8bb5ef",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb8d6e0ce50>"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the results of the monetary model\n",
    "plt.scatter(np.log(in_sample.loc[z].revenue_2015), amount_model_fit.fittedvalues)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bdcdeab9",
   "metadata": {},
   "source": [
    "To summarize we have just calibrated two models the first one over here, to predict the likelihood that someone will be active. And the second one over here to predict how much they will spend if they will be active in 2015. \n",
    "\n",
    "Now the end game of this exercise is actually to apply the models to predict the future. So what we are going to do is to look at today's behavior, today's data, and extract exactly the same information about today's customers as we used in terms of predictors about a year ago. So we're going to extract recency, first purchase, average amount, maximum amount, and everything else for the 2015 customers about whom of course we have no idea who will be active next year, and how much they'll spend in 2016. But we can try to predict that, with our model, so once you create that and that what is usually called the out of sample data set."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d4c72d87",
   "metadata": {},
   "source": [
    "## Apply The Model To Today's Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "4c6e9017",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compute RFM as of Today\n",
    "q = \"\"\"\n",
    "        SELECT customer_id, \n",
    "        MIN(days_since) AS 'recency', \n",
    "        MAX(days_since) AS 'first_purchase', \n",
    "        COUNT(*) AS 'frequency', \n",
    "        AVG(purchase_amount) AS 'avg_amount', \n",
    "        MAX(purchase_amount) AS 'max_amount' FROM df GROUP BY 1\n",
    "\"\"\"\n",
    "customers_2015 = sqldf(q, globals())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "e0bc06d7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predict the target variables based on today's data\n",
    "# 1st: we predict the probability that a customer is going to be in 2016 active\n",
    "customers_2015[\"prob_predicted\"] = prob_model_fit.predict(customers_2015)\n",
    "\n",
    "# 2nd we predict the revenue\n",
    "# Since our model predicts the log(amount) we need to take the exponential of\n",
    "# the predicted revenue generated per customer\n",
    "customers_2015[\"revenue_predicted\"] = np.exp(amount_model_fit.predict(customers_2015))\n",
    "\n",
    "# 3rd: we give a score to each customer which is the conjunction of probability\n",
    "# predicted multiplied by the amount.\n",
    "customers_2015[\"score_predicted\"] = (\n",
    "    customers_2015[\"prob_predicted\"] * customers_2015[\"revenue_predicted\"]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "fb9e20e6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count   18,417.00\n",
       "mean         0.22\n",
       "std          0.25\n",
       "min          0.00\n",
       "25%          0.01\n",
       "50%          0.11\n",
       "75%          0.40\n",
       "max          1.00\n",
       "Name: prob_predicted, dtype: float64"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_2015.prob_predicted.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "a8d0cbae",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count   18,417.00\n",
       "mean        65.63\n",
       "std        147.89\n",
       "min          6.54\n",
       "25%         29.00\n",
       "50%         35.05\n",
       "75%         57.30\n",
       "max      3,832.95\n",
       "Name: revenue_predicted, dtype: float64"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_2015.revenue_predicted.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "685c1b9d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count   18,417.00\n",
       "mean        18.83\n",
       "std         70.21\n",
       "min          0.00\n",
       "25%          0.46\n",
       "50%          4.56\n",
       "75%         17.96\n",
       "max      2,854.16\n",
       "Name: score_predicted, dtype: float64"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_2015.score_predicted.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c844bacc",
   "metadata": {},
   "source": [
    "So the overall score Is actually a function of both and probability and revenue together. And the score has a mean of 18,8. What does it mean? From a managerial point of view, that value is extremely important. It means that, on the average, every customer in this database, will spend 18 point eight dollar next year."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "766ed6aa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "customers_2015.score_predicted.hist(bins=30)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e898643",
   "metadata": {},
   "source": [
    "We'll take the customers here. Look at their predict score and only retain those with the score. Above \\\\$50, and if you look at that, it will create a vector of people with a score above \\\\$50, which contains a total of 1323 customers. So in the list of 18,000 customers here about 1,300 have a predicted score of \\\\$50 or more and you can see which ones they are."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "a68e10a1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1324\n"
     ]
    }
   ],
   "source": [
    "# How many customers have an expected revenue of more than $50\n",
    "z = customers_2015[customers_2015.score_predicted > 50].index.tolist()\n",
    "print(len(z))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "a7c2e8f9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2            90\n",
       "17          480\n",
       "38         1020\n",
       "41         1050\n",
       "42         1150\n",
       "          ...  \n",
       "18372    263080\n",
       "18391    263340\n",
       "18392    263350\n",
       "18402    263470\n",
       "18416    264200\n",
       "Name: customer_id, Length: 1324, dtype: int64"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Customers with the highest score:\n",
    "customers_2015[\"customer_id\"].loc[z]"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}