1_eda.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "HOipZqpEGd8C"
   },
   "source": [
    "# Problem Statement\n",
    "\n",
    "Build a Recommendation System, that can recommend Technology Domains & Tags for questions posted on Stackoverflow.\n",
    "Given a question in Stack-Overflow, predict the **Technology Domain & Associated Tags** for it.\n",
    "\n",
    "**Note** : Here we are working with a only 14 Technology Domain & less than 10% data from Stackoverflow for demonstration purpose only."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "GiwHQ5IVF04A"
   },
   "source": [
    "# Exploratory Data Analysis\n",
    "\n",
    "As the rule of thumb we first check and explore our data to see how does it look like in terms of number.\n",
    "\n",
    "We have the following raw data files :\n",
    "\n",
    "  1. stack_network_nodes.csv** : This file lists a small subset of tags that we focus on along we their *Groups*. We have named the groups for more human understandable form for now.\n",
    "  2. Questions.csv ** : This file contains the questions & related metadata for the questions from Stackoverflow\n",
    "  3. Tags.csv ** : This file contains the Tags to the Questions in the previous file and are mapped via unique IDs.\n",
    "\n",
    "\n",
    "  Data Source :  \n",
    "  1.  10% Data from Stackoverflow : https://www.kaggle.com/stackoverflow/stacksample\n",
    "  2.  Tags used for Demonstrations : https://www.kaggle.com/stackoverflow/stack-overflow-tag-network\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "jahEkEcJHYCU",
    "outputId": "f299314e-ae2b-4a36-8a2e-2e6bf666e5ca"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
     ]
    }
   ],
   "source": [
    "# Let's mount our G-Drive. \n",
    "\n",
    "from google.colab import drive\n",
    "drive.mount('/content/drive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "lx5ZD68GSxcl"
   },
   "outputs": [],
   "source": [
    "# Let's import our packages !\n",
    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "B7wUIBbCSxcq"
   },
   "source": [
    "## Data Loader\n",
    "\n",
    "1. Load the CSV files with the right encoding\n",
    "2. Don't forget to check the shape & head of the input files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "pmOERYH8KrE4"
   },
   "outputs": [],
   "source": [
    "path = '/content/drive/My Drive/ICDMAI_Tutorial'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dpQGeNyZSxcr"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 159
    },
    "colab_type": "code",
    "id": "Esbzg_WiSxcu",
    "outputId": "99e6ecb1-3bff-4574-e612-db15fb69934b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Questions CSV Shape :(1264216, 7) \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Id</th>\n",
       "      <th>OwnerUserId</th>\n",
       "      <th>CreationDate</th>\n",
       "      <th>ClosedDate</th>\n",
       "      <th>Score</th>\n",
       "      <th>Title</th>\n",
       "      <th>Body</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>80</td>\n",
       "      <td>26.0</td>\n",
       "      <td>2008-08-01T13:57:07Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>26</td>\n",
       "      <td>SQLStatement.execute() - multiple queries in o...</td>\n",
       "      <td>&lt;p&gt;I've written a database generation script i...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>90</td>\n",
       "      <td>58.0</td>\n",
       "      <td>2008-08-01T14:41:24Z</td>\n",
       "      <td>2012-12-26T03:45:49Z</td>\n",
       "      <td>144</td>\n",
       "      <td>Good branching and merging tutorials for Torto...</td>\n",
       "      <td>&lt;p&gt;Are there any really good tutorials explain...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>120</td>\n",
       "      <td>83.0</td>\n",
       "      <td>2008-08-01T15:50:08Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>21</td>\n",
       "      <td>ASP.NET Site Maps</td>\n",
       "      <td>&lt;p&gt;Has anyone got experience creating &lt;strong&gt;...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Id  ...                                               Body\n",
       "0   80  ...  <p>I've written a database generation script i...\n",
       "1   90  ...  <p>Are there any really good tutorials explain...\n",
       "2  120  ...  <p>Has anyone got experience creating <strong>...\n",
       "\n",
       "[3 rows x 7 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load Stackoverflow Questions\n",
    "questions_path  = path + '/stacksample/Questions.csv'\n",
    "question = pd.read_csv(questions_path,encoding='latin-1',engine='python')\n",
    "\n",
    "print(\"Questions CSV Shape :{} \".format(question.shape))\n",
    "question.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 153
    },
    "colab_type": "code",
    "id": "ZOqyhb1cbZNf",
    "outputId": "de1929a0-8144-415c-e3ae-ec62441759e4"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Id                int64\n",
       "OwnerUserId     float64\n",
       "CreationDate     object\n",
       "ClosedDate       object\n",
       "Score             int64\n",
       "Title            object\n",
       "Body             object\n",
       "dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking the data types \n",
    "question.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "colab_type": "code",
    "id": "S2k152z8Sxcx",
    "outputId": "915621b9-75c5-4211-9c72-b1e0ceeec840"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tags CSV Shape :(3750994, 2) \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Id</th>\n",
       "      <th>Tag</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>80</td>\n",
       "      <td>flex</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>80</td>\n",
       "      <td>actionscript-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>80</td>\n",
       "      <td>air</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>90</td>\n",
       "      <td>svn</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>90</td>\n",
       "      <td>tortoisesvn</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Id             Tag\n",
       "0  80            flex\n",
       "1  80  actionscript-3\n",
       "2  80             air\n",
       "3  90             svn\n",
       "4  90     tortoisesvn"
      ]
     },
     "execution_count": 6,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load the Tags for the Question IDs in the question\n",
    "question_tag_path = path  + '/stacksample/Tags.csv'\n",
    "\n",
    "question_tag = pd.read_csv(question_tag_path,encoding='latin-1')\n",
    "print(\"Tags CSV Shape :{} \".format(question_tag.shape))\n",
    "question_tag.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "colab_type": "code",
    "id": "m_x9qd9gduu4",
    "outputId": "ebf79542-2246-4a85-a0a6-7e53972e86f9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tag-Group Mapping CSV Shape :(115, 4) \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name</th>\n",
       "      <th>group</th>\n",
       "      <th>nodesize</th>\n",
       "      <th>group_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>c</td>\n",
       "      <td>1</td>\n",
       "      <td>189.83</td>\n",
       "      <td>Programming</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>c++</td>\n",
       "      <td>1</td>\n",
       "      <td>268.11</td>\n",
       "      <td>Programming</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>django</td>\n",
       "      <td>1</td>\n",
       "      <td>40.91</td>\n",
       "      <td>Programming</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>python</td>\n",
       "      <td>1</td>\n",
       "      <td>438.67</td>\n",
       "      <td>Programming</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>flask</td>\n",
       "      <td>1</td>\n",
       "      <td>9.39</td>\n",
       "      <td>Programming</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     name  group  nodesize   group_name\n",
       "0       c      1    189.83  Programming\n",
       "1     c++      1    268.11  Programming\n",
       "2  django      1     40.91  Programming\n",
       "3  python      1    438.67  Programming\n",
       "4   flask      1      9.39  Programming"
      ]
     },
     "execution_count": 7,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load the Tags & Group relationship\n",
    "tag_grp_path = path + '/stack-overflow-tag-network/stack_network_nodes.csv'\n",
    "tag_grp_id = pd.read_csv(tag_grp_path)\n",
    "\n",
    "print(\"tag-Group Mapping CSV Shape :{} \".format(tag_grp_id.shape))\n",
    "tag_grp_id.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Xsi8dhs7Sxc2"
   },
   "source": [
    "# Data Exploration\n",
    "\n",
    "We will perform a number of quick explorations and massage the data for our use-case.For the Demonstration, we are using a fairly smaller dataset of Stackoverflow. But yet it's not that small (~1M questions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "H9kYKdAXdNy3"
   },
   "source": [
    "### 1. Checking  occurence of unique of Tags\n",
    "\n",
    "In 10% sample from Stackoverflow that we have it's shocking to see that we have ~37K Tags.\n",
    "\n",
    "### Things to Wonder ?\n",
    "* Are there spelling mistake in Tags ?\n",
    "* Can javascript or java be tagged as something else too by mistake ?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "colab_type": "code",
    "id": "TfI8d74iSxc8",
    "outputId": "e0aea5a7-8972-48aa-eb66-f370379c7c4d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape : (37034, 2)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tag</th>\n",
       "      <th>occurence_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>16121</th>\n",
       "      <td>javascript</td>\n",
       "      <td>124155</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16020</th>\n",
       "      <td>java</td>\n",
       "      <td>115212</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4301</th>\n",
       "      <td>c#</td>\n",
       "      <td>101186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24211</th>\n",
       "      <td>php</td>\n",
       "      <td>98808</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>979</th>\n",
       "      <td>android</td>\n",
       "      <td>90659</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Tag  occurence_count\n",
       "16121  javascript           124155\n",
       "16020        java           115212\n",
       "4301           c#           101186\n",
       "24211         php            98808\n",
       "979       android            90659"
      ]
     },
     "execution_count": 8,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouping Tags to check frequency of Occurence in the Tags \n",
    "\n",
    "tag_group = question_tag.groupby(by='Tag', as_index=False).agg({'Id': \"count\"})\n",
    "tag_group = tag_group.rename(columns = {'Id': \"occurence_count\"})\n",
    "tag_group = tag_group.sort_values(by=['occurence_count'],ascending=False)\n",
    "print(\"Shape : {}\".format(tag_group.shape))\n",
    "\n",
    "tag_group.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "BENnmv29dtfC"
   },
   "source": [
    "We filter out the tags to only keep Tags for which we have defined Technology Domain. This leaves us with 112 Tags for our dataset only."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "colab_type": "code",
    "id": "Ox88EO4TSxc_",
    "outputId": "0073994f-6453-4e13-81b7-026020c1efe2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape : (112, 5)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tag</th>\n",
       "      <th>occurence_count</th>\n",
       "      <th>group</th>\n",
       "      <th>nodesize</th>\n",
       "      <th>group_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>javascript</td>\n",
       "      <td>124155</td>\n",
       "      <td>6</td>\n",
       "      <td>649.16</td>\n",
       "      <td>Front-end/Designing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>java</td>\n",
       "      <td>115212</td>\n",
       "      <td>8</td>\n",
       "      <td>610.65</td>\n",
       "      <td>MVC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>c#</td>\n",
       "      <td>101186</td>\n",
       "      <td>2</td>\n",
       "      <td>321.13</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>php</td>\n",
       "      <td>98808</td>\n",
       "      <td>6</td>\n",
       "      <td>361.22</td>\n",
       "      <td>Front-end/Designing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>android</td>\n",
       "      <td>90659</td>\n",
       "      <td>4</td>\n",
       "      <td>229.86</td>\n",
       "      <td>Mobile App Development</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Tag  occurence_count  group  nodesize                  group_name\n",
       "0  javascript           124155      6    649.16         Front-end/Designing\n",
       "1        java           115212      8    610.65                         MVC\n",
       "2          c#           101186      2    321.13  MS-Development Environment\n",
       "3         php            98808      6    361.22         Front-end/Designing\n",
       "4     android            90659      4    229.86      Mobile App Development"
      ]
     },
     "execution_count": 9,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merging to get Groups\n",
    "\n",
    "question_tag_grp =  pd.merge(tag_group,tag_grp_id, left_on='Tag',right_on='name')\n",
    "question_tag_grp.pop('name')\n",
    "print(\"Shape : {}\".format(question_tag_grp.shape))\n",
    "\n",
    "question_tag_grp.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Wm-8PB3neOtn"
   },
   "source": [
    "### 2. Check Occurence Ratio\n",
    "We evaluate the presence of a single Tag in the entire Dataset when compared with all the tags.\n",
    "\n",
    "It seems even the most occuring Tag like **javascript** or **java** has just a mere 7% occurence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "colab_type": "code",
    "id": "44ZMrH8cSxdC",
    "outputId": "0b0483ab-a353-4696-b6b9-1f2d0fd620e6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total Occurence Sum : 1621335 \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Tag</th>\n",
       "      <th>occurence_count</th>\n",
       "      <th>group</th>\n",
       "      <th>nodesize</th>\n",
       "      <th>group_name</th>\n",
       "      <th>occurence_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>javascript</td>\n",
       "      <td>124155</td>\n",
       "      <td>6</td>\n",
       "      <td>649.16</td>\n",
       "      <td>Front-end/Designing</td>\n",
       "      <td>0.076576</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>java</td>\n",
       "      <td>115212</td>\n",
       "      <td>8</td>\n",
       "      <td>610.65</td>\n",
       "      <td>MVC</td>\n",
       "      <td>0.071060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>c#</td>\n",
       "      <td>101186</td>\n",
       "      <td>2</td>\n",
       "      <td>321.13</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.062409</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>php</td>\n",
       "      <td>98808</td>\n",
       "      <td>6</td>\n",
       "      <td>361.22</td>\n",
       "      <td>Front-end/Designing</td>\n",
       "      <td>0.060942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>android</td>\n",
       "      <td>90659</td>\n",
       "      <td>4</td>\n",
       "      <td>229.86</td>\n",
       "      <td>Mobile App Development</td>\n",
       "      <td>0.055916</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Tag  occurence_count  ...                  group_name  occurence_ratio\n",
       "0  javascript           124155  ...         Front-end/Designing         0.076576\n",
       "1        java           115212  ...                         MVC         0.071060\n",
       "2          c#           101186  ...  MS-Development Environment         0.062409\n",
       "3         php            98808  ...         Front-end/Designing         0.060942\n",
       "4     android            90659  ...      Mobile App Development         0.055916\n",
       "\n",
       "[5 rows x 6 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Total Tags Occerence \n",
    "occurence_sum = question_tag_grp['occurence_count'].sum()\n",
    "print(\"Total Occurence Sum : {} \".format(occurence_sum))\n",
    "\n",
    "# Evaluating % presence\n",
    "question_tag_grp['occurence_ratio'] =  question_tag_grp['occurence_count'].apply(lambda x : x/occurence_sum)\n",
    "question_tag_grp.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "KiTQtuUeSxdE"
   },
   "source": [
    "### 3. Filtering Questions & Tags\n",
    "\n",
    "Based on the Tags we have for the defined groups, we filter the Questions & related tags from the dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 272
    },
    "colab_type": "code",
    "id": "6CXBEwG7SxdF",
    "outputId": "3d6e8f97-b11b-459b-b49f-97bc2dfdca7c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initial # of Question-Tag mapping with 37k tags : (3750994, 2)\n",
      "Filtered # of Question-Tag mapping with 112 Tags : (1621335, 7)\n",
      "Reduction : 2129659\n",
      "Reduction of  56.776 % in the dataset density due to removal of Tags\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Id</th>\n",
       "      <th>Tag</th>\n",
       "      <th>occurence_count</th>\n",
       "      <th>group</th>\n",
       "      <th>nodesize</th>\n",
       "      <th>group_name</th>\n",
       "      <th>occurence_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1970</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2120</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2840</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4110</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Id  Tag  ...                  group_name  occurence_ratio\n",
       "0   120  sql  ...  MS-Development Environment         0.022069\n",
       "1  1970  sql  ...  MS-Development Environment         0.022069\n",
       "2  2120  sql  ...  MS-Development Environment         0.022069\n",
       "3  2840  sql  ...  MS-Development Environment         0.022069\n",
       "4  4110  sql  ...  MS-Development Environment         0.022069\n",
       "\n",
       "[5 rows x 7 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Filtering Tags\n",
    "\n",
    "print(\"Initial # of Question-Tag mapping with 37k tags : {}\".format(question_tag.shape))\n",
    "\n",
    "filtered_question_tag =  pd.merge(question_tag,question_tag_grp,left_on='Tag',right_on='Tag')\n",
    "print(\"Filtered # of Question-Tag mapping with 112 Tags : {}\".format(filtered_question_tag.shape))\n",
    "\n",
    "decrease =  question_tag.shape[0] -filtered_question_tag.shape[0]\n",
    "print(\"Reduction : {}\".format(decrease))\n",
    "print(\"Reduction of {0: .3f} % in the dataset density due to removal of Tags\".format((decrease/question_tag.shape[0]) * 100 ))\n",
    "\n",
    "filtered_question_tag.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 238
    },
    "colab_type": "code",
    "id": "F68OluamSxdH",
    "outputId": "36081e06-fdd6-44c4-b88d-af34840c297d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Filtered # of Tag occurence for Training : 1621335\n",
      "Grouped  # of Tag occurence for Training: 1051992\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Id</th>\n",
       "      <th>Tag</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120</td>\n",
       "      <td>[sql, asp.net]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>260</td>\n",
       "      <td>[c#, .net]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>330</td>\n",
       "      <td>[c++]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>470</td>\n",
       "      <td>[web-services, .net]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>580</td>\n",
       "      <td>[sql-server]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Id                   Tag\n",
       "0  120        [sql, asp.net]\n",
       "1  260            [c#, .net]\n",
       "2  330                 [c++]\n",
       "3  470  [web-services, .net]\n",
       "4  580          [sql-server]"
      ]
     },
     "execution_count": 12,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouping Tags on Document Level\n",
    "# As 1 Document has multiple tags\n",
    "\n",
    "question_tag_mapping = filtered_question_tag.groupby(by='Id', as_index=False).agg({'Tag':lambda x: list(set(x))})\n",
    "\n",
    "print(\"Filtered # of Tag occurence for Training : {}\".format(filtered_question_tag.shape[0]))\n",
    "print(\"Grouped  # of Tag occurence for Training: {}\".format(question_tag_mapping.shape[0]))\n",
    "\n",
    "question_tag_mapping.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 210
    },
    "colab_type": "code",
    "id": "CrEOvGskSxdK",
    "outputId": "ab2f2339-7066-445b-e917-1ebd4bf270ce"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original # of Questions for Training : 1264216\n",
      "Unique Questions for Training after filtering : 1051992\n",
      "Reduction in Dataset  : 212224\n",
      "Reduction of 16.787 % in Questions Dataset\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Id</th>\n",
       "      <th>OwnerUserId</th>\n",
       "      <th>CreationDate</th>\n",
       "      <th>ClosedDate</th>\n",
       "      <th>Score</th>\n",
       "      <th>Title</th>\n",
       "      <th>Body</th>\n",
       "      <th>Tag</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120</td>\n",
       "      <td>83.0</td>\n",
       "      <td>2008-08-01T15:50:08Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>21</td>\n",
       "      <td>ASP.NET Site Maps</td>\n",
       "      <td>&lt;p&gt;Has anyone got experience creating &lt;strong&gt;...</td>\n",
       "      <td>[sql, asp.net]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>260</td>\n",
       "      <td>91.0</td>\n",
       "      <td>2008-08-01T23:22:08Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>49</td>\n",
       "      <td>Adding scripting functionality to .NET applica...</td>\n",
       "      <td>&lt;p&gt;I have a little game written in C#. It uses...</td>\n",
       "      <td>[c#, .net]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>330</td>\n",
       "      <td>63.0</td>\n",
       "      <td>2008-08-02T02:51:36Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>29</td>\n",
       "      <td>Should I use nested classes in this case?</td>\n",
       "      <td>&lt;p&gt;I am working on a collection of classes use...</td>\n",
       "      <td>[c++]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Id  ...             Tag\n",
       "0  120  ...  [sql, asp.net]\n",
       "1  260  ...      [c#, .net]\n",
       "2  330  ...           [c++]\n",
       "\n",
       "[3 rows x 8 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Filtering Questions based on the Groups of Tags & Merging them\n",
    "\n",
    "filtered_questions =  pd.merge(question,question_tag_mapping)\n",
    "\n",
    "print(\"Original # of Questions for Training : {}\".format(question.shape[0]))\n",
    "print(\"Unique Questions for Training after filtering : {}\".format(filtered_questions.shape[0]))\n",
    "print(\"Reduction in Dataset  : {}\".format(question.shape[0]-filtered_questions.shape[0]))\n",
    " \n",
    "reduction_percentage = ((question.shape[0] - filtered_questions.shape[0])/question.shape[0]) * 100\n",
    "print(\"Reduction of {0:.3f} % in Questions Dataset\".format(reduction_percentage))\n",
    "\n",
    "filtered_questions.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "vDzlAEIUkM35"
   },
   "source": [
    "#### 4.  Checking Distribution in Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 452
    },
    "colab_type": "code",
    "id": "4GakGQKFSxdM",
    "outputId": "b5d7969d-eaaf-4ce2-b6b7-506f6d518fae"
   },
   "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>Id</th>\n",
       "      <th>Tag</th>\n",
       "      <th>tag_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120</td>\n",
       "      <td>[sql, asp.net]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>260</td>\n",
       "      <td>[c#, .net]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>330</td>\n",
       "      <td>[c++]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>470</td>\n",
       "      <td>[web-services, .net]</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>580</td>\n",
       "      <td>[sql-server]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Id                   Tag  tag_count\n",
       "0  120        [sql, asp.net]          2\n",
       "1  260            [c#, .net]          2\n",
       "2  330                 [c++]          1\n",
       "3  470  [web-services, .net]          2\n",
       "4  580          [sql-server]          1"
      ]
     },
     "execution_count": 14,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAD4CAYAAAAZ1BptAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMiQkSVX/H3VXFtgDmrssAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Checking Tags Count Distribution on Document Level\n",
    "\n",
    "question_tag_mapping['tag_count'] = question_tag_mapping['Tag'].apply(lambda x : len(x))\n",
    "question_tag_mapping['tag_count'].hist()\n",
    "question_tag_mapping.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Aa63sy5s_1G1"
   },
   "source": [
    "# Merged dataset prepared for model training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 289
    },
    "colab_type": "code",
    "id": "NSJQX2gwABCk",
    "outputId": "f51d6eec-a74c-4123-8494-9d23fd00fa2f"
   },
   "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>Id</th>\n",
       "      <th>OwnerUserId</th>\n",
       "      <th>CreationDate</th>\n",
       "      <th>ClosedDate</th>\n",
       "      <th>Score</th>\n",
       "      <th>Title</th>\n",
       "      <th>Body</th>\n",
       "      <th>Tag</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120</td>\n",
       "      <td>83.0</td>\n",
       "      <td>2008-08-01T15:50:08Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>21</td>\n",
       "      <td>ASP.NET Site Maps</td>\n",
       "      <td>&lt;p&gt;Has anyone got experience creating &lt;strong&gt;...</td>\n",
       "      <td>[sql, asp.net]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>260</td>\n",
       "      <td>91.0</td>\n",
       "      <td>2008-08-01T23:22:08Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>49</td>\n",
       "      <td>Adding scripting functionality to .NET applica...</td>\n",
       "      <td>&lt;p&gt;I have a little game written in C#. It uses...</td>\n",
       "      <td>[c#, .net]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>330</td>\n",
       "      <td>63.0</td>\n",
       "      <td>2008-08-02T02:51:36Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>29</td>\n",
       "      <td>Should I use nested classes in this case?</td>\n",
       "      <td>&lt;p&gt;I am working on a collection of classes use...</td>\n",
       "      <td>[c++]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>470</td>\n",
       "      <td>71.0</td>\n",
       "      <td>2008-08-02T15:11:47Z</td>\n",
       "      <td>2016-03-26T05:23:29Z</td>\n",
       "      <td>13</td>\n",
       "      <td>Homegrown consumption of web services</td>\n",
       "      <td>&lt;p&gt;I've been writing a few web services for a ...</td>\n",
       "      <td>[web-services, .net]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>580</td>\n",
       "      <td>91.0</td>\n",
       "      <td>2008-08-02T23:30:59Z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>21</td>\n",
       "      <td>Deploying SQL Server Databases from Test to Live</td>\n",
       "      <td>&lt;p&gt;I wonder how you guys manage deployment of ...</td>\n",
       "      <td>[sql-server]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Id  ...                   Tag\n",
       "0  120  ...        [sql, asp.net]\n",
       "1  260  ...            [c#, .net]\n",
       "2  330  ...                 [c++]\n",
       "3  470  ...  [web-services, .net]\n",
       "4  580  ...          [sql-server]\n",
       "\n",
       "[5 rows x 8 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_questions.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 204
    },
    "colab_type": "code",
    "id": "EFeNKb27SxdR",
    "outputId": "7337b545-80f1-4a47-a645-ac7004d11a05"
   },
   "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>Id</th>\n",
       "      <th>Tag</th>\n",
       "      <th>occurence_count</th>\n",
       "      <th>group</th>\n",
       "      <th>nodesize</th>\n",
       "      <th>group_name</th>\n",
       "      <th>occurence_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1970</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2120</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2840</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4110</td>\n",
       "      <td>sql</td>\n",
       "      <td>35782</td>\n",
       "      <td>2</td>\n",
       "      <td>154.23</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>0.022069</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Id  Tag  ...                  group_name  occurence_ratio\n",
       "0   120  sql  ...  MS-Development Environment         0.022069\n",
       "1  1970  sql  ...  MS-Development Environment         0.022069\n",
       "2  2120  sql  ...  MS-Development Environment         0.022069\n",
       "3  2840  sql  ...  MS-Development Environment         0.022069\n",
       "4  4110  sql  ...  MS-Development Environment         0.022069\n",
       "\n",
       "[5 rows x 7 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_question_tag.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 452
    },
    "colab_type": "code",
    "id": "qvTibTJ4SxdY",
    "outputId": "efd1a39d-07fe-4d50-e707-45090c2d2737"
   },
   "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>Tag</th>\n",
       "      <th>occurence_count</th>\n",
       "      <th>group</th>\n",
       "      <th>nodesize</th>\n",
       "      <th>group_name</th>\n",
       "      <th>occurence_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>javascript</td>\n",
       "      <td>124155</td>\n",
       "      <td>6</td>\n",
       "      <td>649.16</td>\n",
       "      <td>Front-end/Designing</td>\n",
       "      <td>7.657578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>java</td>\n",
       "      <td>115212</td>\n",
       "      <td>8</td>\n",
       "      <td>610.65</td>\n",
       "      <td>MVC</td>\n",
       "      <td>7.105996</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>c#</td>\n",
       "      <td>101186</td>\n",
       "      <td>2</td>\n",
       "      <td>321.13</td>\n",
       "      <td>MS-Development Environment</td>\n",
       "      <td>6.240906</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>php</td>\n",
       "      <td>98808</td>\n",
       "      <td>6</td>\n",
       "      <td>361.22</td>\n",
       "      <td>Front-end/Designing</td>\n",
       "      <td>6.094237</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>android</td>\n",
       "      <td>90659</td>\n",
       "      <td>4</td>\n",
       "      <td>229.86</td>\n",
       "      <td>Mobile App Development</td>\n",
       "      <td>5.591627</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Tag  occurence_count  ...                  group_name  occurence_ratio\n",
       "0  javascript           124155  ...         Front-end/Designing         7.657578\n",
       "1        java           115212  ...                         MVC         7.105996\n",
       "2          c#           101186  ...  MS-Development Environment         6.240906\n",
       "3         php            98808  ...         Front-end/Designing         6.094237\n",
       "4     android            90659  ...      Mobile App Development         5.591627\n",
       "\n",
       "[5 rows x 6 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAD4CAYAAADxeG0DAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "question_tag_grp['occurence_ratio'] = question_tag_grp['occurence_ratio'] * 100\n",
    "question_tag_grp['occurence_ratio'].hist(bins=10)\n",
    "question_tag_grp.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 221
    },
    "colab_type": "code",
    "id": "jP0GZCPzSxdi",
    "outputId": "6b451665-ef2d-4285-97e2-ba092e40f67c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unique Questions for Training : (1051992, 2)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Id</th>\n",
       "      <th>group</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120</td>\n",
       "      <td>[2, 2]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>260</td>\n",
       "      <td>[2, 2]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>330</td>\n",
       "      <td>[1]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>470</td>\n",
       "      <td>[2, 8]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>580</td>\n",
       "      <td>[2]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Id   group\n",
       "0  120  [2, 2]\n",
       "1  260  [2, 2]\n",
       "2  330     [1]\n",
       "3  470  [2, 8]\n",
       "4  580     [2]"
      ]
     },
     "execution_count": 18,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouping Tags-Group for a Single Document\n",
    "\n",
    "question_group_mapping = filtered_question_tag.groupby(by='Id', as_index=False).agg({'group':lambda x: list(x)})\n",
    "\n",
    "print(\"Unique Questions for Training : {}\".format(question_group_mapping.shape))\n",
    "question_group_mapping.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 452
    },
    "colab_type": "code",
    "id": "QRPfWGoOSxdl",
    "outputId": "2974162b-a424-4bb5-da6e-5f919ac37104"
   },
   "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>Id</th>\n",
       "      <th>group</th>\n",
       "      <th>unique_grp_count</th>\n",
       "      <th>grp_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>120</td>\n",
       "      <td>[2, 2]</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>260</td>\n",
       "      <td>[2, 2]</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>330</td>\n",
       "      <td>[1]</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>470</td>\n",
       "      <td>[2, 8]</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>580</td>\n",
       "      <td>[2]</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Id   group  unique_grp_count  grp_count\n",
       "0  120  [2, 2]                 1          2\n",
       "1  260  [2, 2]                 1          2\n",
       "2  330     [1]                 1          1\n",
       "3  470  [2, 8]                 2          2\n",
       "4  580     [2]                 1          1"
      ]
     },
     "execution_count": 19,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAD4CAYAAAAZ1BptAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Checking Group Count Distribution on Document Level for Uniqueness\n",
    "\n",
    "question_group_mapping['unique_grp_count'] = question_group_mapping['group'].apply(lambda x : len(set(x)))\n",
    "question_group_mapping['grp_count'] = question_group_mapping['group'].apply(lambda x : len(x))\n",
    "question_group_mapping['unique_grp_count'].hist()\n",
    "question_group_mapping.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "id": "ju2r_SPsSxdn",
    "outputId": "d3464432-17d8-436a-c19c-ec165a331bfe"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# of Documents with Multiple Tags : 318884\n",
      "30.312 % of Document has Multiple Tags in training data \n"
     ]
    }
   ],
   "source": [
    "# Check % of Documents with Multiple Tags in the Training Data \n",
    "\n",
    "mulitple_tags_count = question_group_mapping[question_group_mapping.grp_count != question_group_mapping.unique_grp_count].shape[0]\n",
    "\n",
    "print(\"# of Documents with Multiple Tags : {}\".format(mulitple_tags_count))\n",
    "print(\"{0:.3f} % of Document has Multiple Tags in training data \".format((mulitple_tags_count/question_group_mapping.shape[0])*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 488
    },
    "colab_type": "code",
    "id": "h7qmrdpsKHZL",
    "outputId": "2f4bbc2f-d8af-4bb2-ad6b-2ceef1ae91e9"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'tag')"
      ]
     },
     "execution_count": 21,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAACQQAAAGyCAYAAAC4WFsnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 2880x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Overall view of tag vs documents count\n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "filtered_questions_copy = filtered_questions\n",
    "\n",
    "from sklearn.preprocessing import MultiLabelBinarizer\n",
    "\n",
    "mlb = MultiLabelBinarizer()\n",
    "filtered_questions_copy['group_Tag'] = filtered_questions_copy['Tag']\n",
    "filtered_questions_copy = filtered_questions_copy.join(pd.DataFrame(mlb.fit_transform(filtered_questions_copy.pop('group_Tag')),\n",
    "                          columns=mlb.classes_,\n",
    "                          index=filtered_questions_copy.index))\n",
    "\n",
    "question_tag_summary = filtered_questions_copy.drop(['Id', 'OwnerUserId', 'CreationDate', 'ClosedDate', 'Score', 'Title',\n",
    "       'Body','Tag'], axis=1)\n",
    "categories = list(question_tag_summary.columns.values)\n",
    "counts = []\n",
    "for i in categories:\n",
    "    counts.append((i, question_tag_summary[i].sum()))\n",
    "df_stats = pd.DataFrame(counts, columns=['category', 'number_of_questions'])\n",
    "\n",
    "df_stats.plot(x='category', y='number_of_questions', kind='bar', legend=False, grid=True, figsize=(40, 5))\n",
    "plt.title(\"Number of documents per tag\")\n",
    "plt.ylabel('# of documents', fontsize=12)\n",
    "plt.xlabel('tag', fontsize=12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 282
    },
    "colab_type": "code",
    "id": "cjuR27X_LbtH",
    "outputId": "73ede6d0-13b8-4f55-bb70-ee3ca7e10576"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f2f6ea4de80>"
      ]
     },
     "execution_count": 22,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAD4CAYAAADsKpHdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Histogram of docuement len\n",
    "lens = filtered_questions_copy.Body.str.len()\n",
    "lens.hist(bins = np.arange(0,5000,50))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "g26ra6abd4cp"
   },
   "source": [
    "### 5. Checking Creation date & it's distribution\n",
    "\n",
    "Though we don't use it, but it shows that December has the least number of questions. Holiday Season much"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 496
    },
    "colab_type": "code",
    "id": "lOnnms9keQdN",
    "outputId": "33d95893-f3fe-48a3-e107-0fffcd1653c2"
   },
   "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>Id</th>\n",
       "      <th>OwnerUserId</th>\n",
       "      <th>CreationDate</th>\n",
       "      <th>ClosedDate</th>\n",
       "      <th>Score</th>\n",
       "      <th>Title</th>\n",
       "      <th>Body</th>\n",
       "      <th>CreationMonth</th>\n",
       "      <th>CreationYear</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>80</td>\n",
       "      <td>26.0</td>\n",
       "      <td>2008-08-01 13:57:07</td>\n",
       "      <td>NaN</td>\n",
       "      <td>26</td>\n",
       "      <td>SQLStatement.execute() - multiple queries in o...</td>\n",
       "      <td>&lt;p&gt;I've written a database generation script i...</td>\n",
       "      <td>8</td>\n",
       "      <td>2008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>90</td>\n",
       "      <td>58.0</td>\n",
       "      <td>2008-08-01 14:41:24</td>\n",
       "      <td>2012-12-26T03:45:49Z</td>\n",
       "      <td>144</td>\n",
       "      <td>Good branching and merging tutorials for Torto...</td>\n",
       "      <td>&lt;p&gt;Are there any really good tutorials explain...</td>\n",
       "      <td>8</td>\n",
       "      <td>2008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>120</td>\n",
       "      <td>83.0</td>\n",
       "      <td>2008-08-01 15:50:08</td>\n",
       "      <td>NaN</td>\n",
       "      <td>21</td>\n",
       "      <td>ASP.NET Site Maps</td>\n",
       "      <td>&lt;p&gt;Has anyone got experience creating &lt;strong&gt;...</td>\n",
       "      <td>8</td>\n",
       "      <td>2008</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Id  OwnerUserId  ... CreationMonth CreationYear\n",
       "0   80         26.0  ...             8         2008\n",
       "1   90         58.0  ...             8         2008\n",
       "2  120         83.0  ...             8         2008\n",
       "\n",
       "[3 rows x 9 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAEvCAYAAAAemFY+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAWQElEQVR4nO3dfYxddZ3H8fd3W9FSlycxE7ZttiQ2\nmErXFSdQl8RMrAvDQyx/KIG4Utyu/UNQNE3csrtJE5UNZkWErLJppLa4hMoiGxop1gaYmE0W5MlQ\nobJMeJxuedDyYCHKjn73j/tDL8MMA3N7z/3NnfcrmfSc7/mdc779NbQfzsO9kZlIkiSpLn/S6wYk\nSZL0eoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmSpArN73UDB9vRRx+dS5cu7XUbVXrppZdY\nuHBhr9uYU5zzZjnfzXK+m+V8N6up+b7nnnt+mZnvnmxb34W0pUuXcvfdd/e6jSqNjIwwNDTU6zbm\nFOe8Wc53s5zvZjnfzWpqviPi8am2ebtTkiSpQoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmS\npAoZ0iRJkipkSJMkSaqQIU2SJKlChjRJkqQKGdIkSZIq1Hff3anZaemGm3vdwpvy2KVn9LoFSdIc\n4ZU0SZKkChnSJEmSKmRIkyRJqpDPpEnSm+Bzk5Ka5pU0SZKkChnSJEmSKuTtTukteKu3vNavGOf8\nHtwm85aXJM1+hjRJ6iNNPjvXyf+E+D8S0vS83SlJklQhQ5okSVKFvN0p9aHZ8nER4G0v1a+W/56m\nu73sf0v9x5AmSWpcLcFHqpm3OyVJkipkSJMkSaqQIU2SJKlChjRJkqQK+eKApJ46mA+Q9+obHiSp\nG7ySJkmSVCFDmiRJUoUMaZIkSRUypEmSJFXIkCZJklQhQ5okSVKFDGmSJEkVMqRJkiRVyJAmSZJU\noWm/cSAiNgNnAs9k5vGldhTwfWAp8BhwdmY+FxEBXAGcDrwMnJ+Z95Z91gD/VA771czcWuofBLYA\nC4AdwEWZmVOdo+Pf8Rwy8ZPc/TR2SZJmjzdzJW0LMDyhtgG4NTOXAbeWdYDTgGXlZx1wFfwh1G0E\nTgJOBDZGxJFln6uAz7TtNzzNOSRJkvretCEtM38C7J9QXg1sLctbgbPa6tdkyx3AERFxDHAqsCsz\n95erYbuA4bLtsMy8IzMTuGbCsSY7hyRJUt+b6TNpA5m5ryw/BQyU5UXAk23jxkrtjepjk9Tf6ByS\nJEl9b9pn0qZTnh/Lg9HMTM8REeto3V5lYGCAkZGRbrYza6xfMf6a9YEFr6+pu5zzZjnfzXK+mzXd\nfPtv38F14MCBns/pTEPa0xFxTGbuK7csnyn1vcCStnGLS20vMDShPlLqiycZ/0bneJ3M3ARsAhgc\nHMyhoaGphh4UEx/Ir9dr/3jXrxjnst0d53K9Bc55s5zvZjnfzZpuvh/75FBzzcwBIyMjdDtPTGem\ntzu3A2vK8hrgprb6edGyEnih3LLcCZwSEUeWFwZOAXaWbS9GxMryZuh5E4412TkkSZL63pv5CI7r\naF0FOzoixmi9pXkpcH1ErAUeB84uw3fQ+viNUVofwfFpgMzcHxFfAe4q476cma++jPBZ/vgRHLeU\nH97gHJIkSX1v2pCWmedOsWnVJGMTuGCK42wGNk9Svxs4fpL6ryY7hyRJ0lzgNw5IkiRVyJAmSZJU\nIUOaJElShQxpkiRJFTKkSZIkVciQJkmSVCFDmiRJUoUMaZIkSRXyS9ckSeoDs+V7pR+79IxetzBr\neCVNkiSpQoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmSpAoZ0iRJkipkSJMkSaqQIU2SJKlC\nhjRJkqQK+bVQkiSpMbPl66u2DC/sdQteSZMkSaqRIU2SJKlChjRJkqQKGdIkSZIqZEiTJEmqkCFN\nkiSpQoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmSpAoZ0iRJkipkSJMkSaqQIU2SJKlChjRJ\nkqQKGdIkSZIqZEiTJEmqUEchLSK+GBEPRMTPI+K6iHhHRBwbEXdGxGhEfD8iDilj317WR8v2pW3H\nubjUH4qIU9vqw6U2GhEbOulVkiRpNplxSIuIRcDngcHMPB6YB5wDfA24PDPfAzwHrC27rAWeK/XL\nyzgiYnnZ733AMPDtiJgXEfOAbwGnAcuBc8tYSZKkvtfp7c75wIKImA8cCuwDPgLcULZvBc4qy6vL\nOmX7qoiIUt+Wmb/NzEeBUeDE8jOamY9k5ivAtjJWkiSp7804pGXmXuDrwBO0wtkLwD3A85k5XoaN\nAYvK8iLgybLveBn/rvb6hH2mqkuSJPW9+TPdMSKOpHVl61jgeeA/aN2ubFxErAPWAQwMDDAyMtLV\n861fMT79oAoNLJi9vc9WznmznO9mOd/Ncr6bdeDAga7nienMOKQBHwUezcxnASLiRuBk4IiImF+u\nli0G9pbxe4ElwFi5PXo48Ku2+qva95mq/hqZuQnYBDA4OJhDQ0Md/Lamd/6Gm7t6/G5Zv2Kcy3Z3\n8keut8o5b5bz3Sznu1nOd7O2DC+k23liOp08k/YEsDIiDi3Plq0CHgRuBz5exqwBbirL28s6Zftt\nmZmlfk55+/NYYBnwU+AuYFl5W/QQWi8XbO+gX0mSpFljxpE8M++MiBuAe4Fx4D5aV7NuBrZFxFdL\n7eqyy9XA9yJiFNhPK3SRmQ9ExPW0At44cEFm/g4gIi4EdtJ6c3RzZj4w034lSZJmk46um2bmRmDj\nhPIjtN7MnDj2N8AnpjjOJcAlk9R3ADs66VGSJGk28hsHJEmSKmRIkyRJqpAhTZIkqUKGNEmSpAoZ\n0iRJkipkSJMkSaqQIU2SJKlChjRJkqQKGdIkSZIqZEiTJEmqkCFNkiSpQoY0SZKkChnSJEmSKmRI\nkyRJqpAhTZIkqUKGNEmSpAoZ0iRJkipkSJMkSaqQIU2SJKlChjRJkqQKGdIkSZIqZEiTJEmqkCFN\nkiSpQoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmSpAoZ0iRJkipkSJMkSaqQIU2SJKlChjRJ\nkqQKGdIkSZIqZEiTJEmqkCFNkiSpQoY0SZKkCnUU0iLiiIi4ISJ+ERF7IuJDEXFUROyKiIfLr0eW\nsRERV0bEaETcHxEntB1nTRn/cESsaat/MCJ2l32ujIjopF9JkqTZotMraVcAP8rM9wLvB/YAG4Bb\nM3MZcGtZBzgNWFZ+1gFXAUTEUcBG4CTgRGDjq8GujPlM237DHfYrSZI0K8w4pEXE4cCHgasBMvOV\nzHweWA1sLcO2AmeV5dXANdlyB3BERBwDnArsysz9mfkcsAsYLtsOy8w7MjOBa9qOJUmS1Nc6uZJ2\nLPAs8N2IuC8ivhMRC4GBzNxXxjwFDJTlRcCTbfuPldob1ccmqUuSJPW9+R3uewLwucy8MyKu4I+3\nNgHIzIyI7KTBNyMi1tG6hcrAwAAjIyNdPd/6FeNdPX63DCyYvb3PVs55s5zvZjnfzXK+m3XgwIGu\n54npdBLSxoCxzLyzrN9AK6Q9HRHHZOa+csvymbJ9L7Ckbf/FpbYXGJpQHyn1xZOMf53M3ARsAhgc\nHMyhoaHJhh0052+4uavH75b1K8a5bHcnf+R6q5zzZjnfzXK+m+V8N2vL8EK6nSemM+PbnZn5FPBk\nRBxXSquAB4HtwKtvaK4BbirL24HzylueK4EXym3RncApEXFkeWHgFGBn2fZiRKwsb3We13YsSZKk\nvtZpJP8ccG1EHAI8AnyaVvC7PiLWAo8DZ5exO4DTgVHg5TKWzNwfEV8B7irjvpyZ+8vyZ4EtwALg\nlvIjSZLU9zoKaZn5M2Bwkk2rJhmbwAVTHGczsHmS+t3A8Z30KEmSNBv5jQOSJEkVMqRJkiRVyJAm\nSZJUIUOaJElShQxpkiRJFTKkSZIkVciQJkmSVCFDmiRJUoUMaZIkSRUypEmSJFXIkCZJklQhQ5ok\nSVKFDGmSJEkVMqRJkiRVyJAmSZJUIUOaJElShQxpkiRJFTKkSZIkVciQJkmSVCFDmiRJUoUMaZIk\nSRUypEmSJFXIkCZJklQhQ5okSVKFDGmSJEkVMqRJkiRVyJAmSZJUIUOaJElShQxpkiRJFTKkSZIk\nVciQJkmSVCFDmiRJUoUMaZIkSRUypEmSJFXIkCZJklQhQ5okSVKFDGmSJEkV6jikRcS8iLgvIn5Y\n1o+NiDsjYjQivh8Rh5T628v6aNm+tO0YF5f6QxFxalt9uNRGI2JDp71KkiTNFgfjStpFwJ629a8B\nl2fme4DngLWlvhZ4rtQvL+OIiOXAOcD7gGHg2yX4zQO+BZwGLAfOLWMlSZL6XkchLSIWA2cA3ynr\nAXwEuKEM2QqcVZZXl3XK9lVl/GpgW2b+NjMfBUaBE8vPaGY+kpmvANvKWEmSpL7X6ZW0bwJfAn5f\n1t8FPJ+Z42V9DFhUlhcBTwKU7S+U8X+oT9hnqrokSVLfmz/THSPiTOCZzLwnIoYOXksz6mUdsA5g\nYGCAkZGRrp5v/Yrx6QdVaGDB7O19tnLOm+V8N8v5bpbz3awDBw50PU9MZ8YhDTgZ+FhEnA68AzgM\nuAI4IiLml6tli4G9ZfxeYAkwFhHzgcOBX7XVX9W+z1T118jMTcAmgMHBwRwaGurgtzW98zfc3NXj\nd8v6FeNctruTP3K9Vc55s5zvZjnfzXK+m7VleCHdzhPTmfHtzsy8ODMXZ+ZSWg/+35aZnwRuBz5e\nhq0BbirL28s6ZfttmZmlfk55+/NYYBnwU+AuYFl5W/SQco7tM+1XkiRpNulGJP97YFtEfBW4D7i6\n1K8GvhcRo8B+WqGLzHwgIq4HHgTGgQsy83cAEXEhsBOYB2zOzAe60K8kSVJ1DkpIy8wRYKQsP0Lr\nzcyJY34DfGKK/S8BLpmkvgPYcTB6lCRJmk38xgFJkqQKGdIkSZIqZEiTJEmqkCFNkiSpQoY0SZKk\nChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmSpAoZ0iRJkipkSJMkSaqQIU2SJKlChjRJkqQKGdIkSZIq\nZEiTJEmqkCFNkiSpQoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmSpAoZ0iRJkipkSJMkSaqQ\nIU2SJKlChjRJkqQKGdIkSZIqZEiTJEmqkCFNkiSpQoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKG\nNEmSpAoZ0iRJkipkSJMkSaqQIU2SJKlCMw5pEbEkIm6PiAcj4oGIuKjUj4qIXRHxcPn1yFKPiLgy\nIkYj4v6IOKHtWGvK+IcjYk1b/YMRsbvsc2VERCe/WUmSpNmikytp48D6zFwOrAQuiIjlwAbg1sxc\nBtxa1gFOA5aVn3XAVdAKdcBG4CTgRGDjq8GujPlM237DHfQrSZI0a8w4pGXmvsy8tyz/GtgDLAJW\nA1vLsK3AWWV5NXBNttwBHBERxwCnArsyc39mPgfsAobLtsMy847MTOCatmNJkiT1tYPyTFpELAU+\nANwJDGTmvrLpKWCgLC8CnmzbbazU3qg+NkldkiSp783v9AAR8U7gB8AXMvPF9sfGMjMjIjs9x5vo\nYR2tW6gMDAwwMjLS1fOtXzHe1eN3y8CC2dv7bOWcN8v5bpbz3Sznu1kHDhzoep6YTkchLSLeRiug\nXZuZN5by0xFxTGbuK7csnyn1vcCStt0Xl9peYGhCfaTUF08y/nUycxOwCWBwcDCHhoYmG3bQnL/h\n5q4ev1vWrxjnst0d53K9Bc55s5zvZjnfzXK+m7VleCHdzhPT6eTtzgCuBvZk5jfaNm0HXn1Dcw1w\nU1v9vPKW50rghXJbdCdwSkQcWV4YOAXYWba9GBEry7nOazuWJElSX+skkp8MfArYHRE/K7V/AC4F\nro+ItcDjwNll2w7gdGAUeBn4NEBm7o+IrwB3lXFfzsz9ZfmzwBZgAXBL+ZEkSep7Mw5pmflfwFSf\nW7ZqkvEJXDDFsTYDmyep3w0cP9MeJUmSZiu/cUCSJKlChjRJkqQKGdIkSZIqZEiTJEmqkCFNkiSp\nQoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmSpAoZ0iRJkipkSJMkSaqQIU2SJKlChjRJkqQK\nGdIkSZIqZEiTJEmqkCFNkiSpQoY0SZKkChnSJEmSKmRIkyRJqpAhTZIkqUKGNEmSpAoZ0iRJkipk\nSJMkSaqQIU2SJKlChjRJkqQKGdIkSZIqZEiTJEmqkCFNkiSpQoY0SZKkChnSJEmSKmRIkyRJqpAh\nTZIkqUKGNEmSpAoZ0iRJkipkSJMkSaqQIU2SJKlC1Ye0iBiOiIciYjQiNvS6H0mSpCZUHdIiYh7w\nLeA0YDlwbkQs721XkiRJ3Vd1SANOBEYz85HMfAXYBqzucU+SJEldV3tIWwQ82bY+VmqSJEl9LTKz\n1z1MKSI+Dgxn5t+V9U8BJ2XmhRPGrQPWldXjgIcabXT2OBr4Za+bmGOc82Y5381yvpvlfDerqfn+\n88x892Qb5jdw8k7sBZa0rS8utdfIzE3Apqaamq0i4u7MHOx1H3OJc94s57tZzneznO9m1TDftd/u\nvAtYFhHHRsQhwDnA9h73JEmS1HVVX0nLzPGIuBDYCcwDNmfmAz1uS5IkqeuqDmkAmbkD2NHrPvqE\nt4Sb55w3y/lulvPdLOe7WT2f76pfHJAkSZqran8mTZIkaU4ypM0BEbEkIm6PiAcj4oGIuKjXPc0F\nETEvIu6LiB/2upd+FxFHRMQNEfGLiNgTER/qdU/9LCK+WP4u+XlEXBcR7+h1T/0mIjZHxDMR8fO2\n2lERsSsiHi6/HtnLHvvJFPP9L+XvlPsj4j8j4oim+zKkzQ3jwPrMXA6sBC7w67UacRGwp9dNzBFX\nAD/KzPcC78d575qIWAR8HhjMzONpvdR1Tm+76ktbgOEJtQ3ArZm5DLi1rOvg2MLr53sXcHxm/gXw\nP8DFTTdlSJsDMnNfZt5bln9N6x8wv7mhiyJiMXAG8J1e99LvIuJw4MPA1QCZ+UpmPt/brvrefGBB\nRMwHDgX+t8f99J3M/Amwf0J5NbC1LG8Fzmq0qT422Xxn5o8zc7ys3kHrs1obZUibYyJiKfAB4M7e\ndtL3vgl8Cfh9rxuZA44FngW+W24vfyciFva6qX6VmXuBrwNPAPuAFzLzx73tas4YyMx9ZfkpYKCX\nzcwxfwvc0vRJDWlzSES8E/gB8IXMfLHX/fSriDgTeCYz7+l1L3PEfOAE4KrM/ADwEt4G6pryHNRq\nWuH4z4CFEfE3ve1q7snWRzP48QwNiIh/pPXY0LVNn9uQNkdExNtoBbRrM/PGXvfT504GPhYRjwHb\ngI9ExL/3tqW+NgaMZearV4dvoBXa1B0fBR7NzGcz8/+AG4G/6nFPc8XTEXEMQPn1mR730/ci4nzg\nTOCT2YPPLDOkzQEREbSe19mTmd/odT/9LjMvzszFmbmU1gPVt2WmVxq6JDOfAp6MiONKaRXwYA9b\n6ndPACsj4tDyd8sqfFGjKduBNWV5DXBTD3vpexExTOuxlY9l5su96MGQNjecDHyK1hWdn5Wf03vd\nlHQQfQ64NiLuB/4S+Oce99O3yhXLG4B7gd20/h3p+Sez95uIuA74b+C4iBiLiLXApcBfR8TDtK5o\nXtrLHvvJFPP9r8CfArvKv5v/1nhffuOAJElSfbySJkmSVCFDmiRJUoUMaZIkSRUypEmSJFXIkCZJ\nklQhQ5okSVKFDGmSJEkVMqRJkiRV6P8B6tnVcBDO+7oAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Changing Object to DateTime format. Also segregating Month & Year\n",
    "\n",
    "question['CreationDate'] = question['CreationDate'].astype('datetime64[ns]') \n",
    "question['CreationMonth'] = question['CreationDate'].apply(lambda x : x.month)\n",
    "question['CreationYear'] = question['CreationDate'].apply(lambda x : x.year)\n",
    "\n",
    "# Checking Distribution of Creation month for Questions\n",
    "fig = plt.figure(figsize = (10,5))\n",
    "ax = fig.gca()\n",
    "question['CreationMonth'].hist(bins=12)\n",
    "\n",
    "question.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "kbODTtboSxd1"
   },
   "source": [
    "### 6. Saving Filtered Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "r9KBnw7PSxd2"
   },
   "outputs": [],
   "source": [
    "filtered_question_tag.head(3)\n",
    "filtered_question_tag.to_pickle('/content/drive/My Drive/ICDMAI_Tutorial/notebook/filtered_data/question_tag_group_mapping.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "eJZZaT4tSxd4"
   },
   "outputs": [],
   "source": [
    "filtered_questions.head(3)\n",
    "filtered_questions.to_pickle('/content/drive/My Drive/ICDMAI_Tutorial/notebook/filtered_data/question_tag_text_mapping.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "XUTeOnAlSxd6"
   },
   "outputs": [],
   "source": [
    "question_tag_mapping.head()\n",
    "question_tag_mapping.to_pickle('/content/drive/My Drive/ICDMAI_Tutorial/notebook/filtered_data/question_tag_id_mapping.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "b09--sIsSxd-"
   },
   "outputs": [],
   "source": [
    "question_group_mapping.head()\n",
    "question_group_mapping.to_pickle('/content/drive/My Drive/ICDMAI_Tutorial/notebook/filtered_data/question_group_mapping.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "KXrg4gXpA1Bd"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "1_eda.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}