06. Pricing Test.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import  warnings\n",
    "warnings.simplefilter('ignore')\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy.stats import ttest_ind\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>user_id</th>\n",
       "      <th>city</th>\n",
       "      <th>country</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>510335</td>\n",
       "      <td>Peabody</td>\n",
       "      <td>USA</td>\n",
       "      <td>42.53</td>\n",
       "      <td>-70.97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>89568</td>\n",
       "      <td>Reno</td>\n",
       "      <td>USA</td>\n",
       "      <td>39.54</td>\n",
       "      <td>-119.82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>434134</td>\n",
       "      <td>Rialto</td>\n",
       "      <td>USA</td>\n",
       "      <td>34.11</td>\n",
       "      <td>-117.39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>289769</td>\n",
       "      <td>Carson City</td>\n",
       "      <td>USA</td>\n",
       "      <td>39.15</td>\n",
       "      <td>-119.74</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>939586</td>\n",
       "      <td>Chicago</td>\n",
       "      <td>USA</td>\n",
       "      <td>41.84</td>\n",
       "      <td>-87.68</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id         city country    lat    long\n",
       "0   510335      Peabody     USA  42.53  -70.97\n",
       "1    89568         Reno     USA  39.54 -119.82\n",
       "2   434134       Rialto     USA  34.11 -117.39\n",
       "3   289769  Carson City     USA  39.15 -119.74\n",
       "4   939586      Chicago     USA  41.84  -87.68"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_table = pd.read_csv('./data/Pricing_Test/user_table.csv')\n",
    "user_table.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 275616 entries, 0 to 275615\n",
      "Data columns (total 5 columns):\n",
      "user_id    275616 non-null int64\n",
      "city       275616 non-null object\n",
      "country    275616 non-null object\n",
      "lat        275616 non-null float64\n",
      "long       275616 non-null float64\n",
      "dtypes: float64(2), int64(1), object(2)\n",
      "memory usage: 10.5+ MB\n"
     ]
    }
   ],
   "source": [
    "user_table.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user_id    0\n",
       "city       0\n",
       "country    0\n",
       "lat        0\n",
       "long       0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user_table.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>user_id</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>source</th>\n",
       "      <th>device</th>\n",
       "      <th>operative_system</th>\n",
       "      <th>test</th>\n",
       "      <th>price</th>\n",
       "      <th>converted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>604839</td>\n",
       "      <td>2015-05-08 03:38:34</td>\n",
       "      <td>ads_facebook</td>\n",
       "      <td>mobile</td>\n",
       "      <td>iOS</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>624057</td>\n",
       "      <td>2015-05-10 21:08:46</td>\n",
       "      <td>seo-google</td>\n",
       "      <td>mobile</td>\n",
       "      <td>android</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>317970</td>\n",
       "      <td>2015-04-04 15:01:23</td>\n",
       "      <td>ads-bing</td>\n",
       "      <td>mobile</td>\n",
       "      <td>android</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>685636</td>\n",
       "      <td>2015-05-07 07:26:01</td>\n",
       "      <td>direct_traffic</td>\n",
       "      <td>mobile</td>\n",
       "      <td>iOS</td>\n",
       "      <td>1</td>\n",
       "      <td>59</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>820854</td>\n",
       "      <td>2015-05-24 11:04:40</td>\n",
       "      <td>ads_facebook</td>\n",
       "      <td>web</td>\n",
       "      <td>mac</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id            timestamp          source  device operative_system  \\\n",
       "0   604839  2015-05-08 03:38:34    ads_facebook  mobile              iOS   \n",
       "1   624057  2015-05-10 21:08:46      seo-google  mobile          android   \n",
       "2   317970  2015-04-04 15:01:23        ads-bing  mobile          android   \n",
       "3   685636  2015-05-07 07:26:01  direct_traffic  mobile              iOS   \n",
       "4   820854  2015-05-24 11:04:40    ads_facebook     web              mac   \n",
       "\n",
       "   test  price  converted  \n",
       "0     0     39          0  \n",
       "1     0     39          0  \n",
       "2     0     39          0  \n",
       "3     1     59          0  \n",
       "4     0     39          0  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results = pd.read_csv('./data/Pricing_Test/test_results.csv', parse_dates=['timestamp'])\n",
    "test_results.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 316800 entries, 0 to 316799\n",
      "Data columns (total 8 columns):\n",
      "user_id             316800 non-null int64\n",
      "timestamp           316800 non-null object\n",
      "source              316800 non-null object\n",
      "device              316800 non-null object\n",
      "operative_system    316800 non-null object\n",
      "test                316800 non-null int64\n",
      "price               316800 non-null int64\n",
      "converted           316800 non-null int64\n",
      "dtypes: int64(4), object(4)\n",
      "memory usage: 19.3+ MB\n"
     ]
    }
   ],
   "source": [
    "test_results.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user_id             0\n",
       "timestamp           0\n",
       "source              0\n",
       "device              0\n",
       "operative_system    0\n",
       "test                0\n",
       "price               0\n",
       "converted           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Processing and Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user_table:\t (275616,)\n",
      "test_results:\t (316800,)\n"
     ]
    }
   ],
   "source": [
    "# check uniquess\n",
    "print('user_table:\\t', user_table['user_id'].unique().shape)\n",
    "print('test_results:\\t', test_results['user_id'].unique().shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>user_id</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>source</th>\n",
       "      <th>device</th>\n",
       "      <th>operative_system</th>\n",
       "      <th>test</th>\n",
       "      <th>price</th>\n",
       "      <th>converted</th>\n",
       "      <th>city</th>\n",
       "      <th>country</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>604839</td>\n",
       "      <td>2015-05-08 03:38:34</td>\n",
       "      <td>ads_facebook</td>\n",
       "      <td>mobile</td>\n",
       "      <td>iOS</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>Buffalo</td>\n",
       "      <td>USA</td>\n",
       "      <td>42.89</td>\n",
       "      <td>-78.86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>624057</td>\n",
       "      <td>2015-05-10 21:08:46</td>\n",
       "      <td>seo-google</td>\n",
       "      <td>mobile</td>\n",
       "      <td>android</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>Lakeville</td>\n",
       "      <td>USA</td>\n",
       "      <td>44.68</td>\n",
       "      <td>-93.24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>317970</td>\n",
       "      <td>2015-04-04 15:01:23</td>\n",
       "      <td>ads-bing</td>\n",
       "      <td>mobile</td>\n",
       "      <td>android</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>Parma</td>\n",
       "      <td>USA</td>\n",
       "      <td>41.38</td>\n",
       "      <td>-81.73</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>685636</td>\n",
       "      <td>2015-05-07 07:26:01</td>\n",
       "      <td>direct_traffic</td>\n",
       "      <td>mobile</td>\n",
       "      <td>iOS</td>\n",
       "      <td>1</td>\n",
       "      <td>59</td>\n",
       "      <td>0</td>\n",
       "      <td>Fayetteville</td>\n",
       "      <td>USA</td>\n",
       "      <td>35.07</td>\n",
       "      <td>-78.90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>820854</td>\n",
       "      <td>2015-05-24 11:04:40</td>\n",
       "      <td>ads_facebook</td>\n",
       "      <td>web</td>\n",
       "      <td>mac</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>Fishers</td>\n",
       "      <td>USA</td>\n",
       "      <td>39.95</td>\n",
       "      <td>-86.02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id            timestamp          source  device operative_system  \\\n",
       "0   604839  2015-05-08 03:38:34    ads_facebook  mobile              iOS   \n",
       "1   624057  2015-05-10 21:08:46      seo-google  mobile          android   \n",
       "2   317970  2015-04-04 15:01:23        ads-bing  mobile          android   \n",
       "3   685636  2015-05-07 07:26:01  direct_traffic  mobile              iOS   \n",
       "4   820854  2015-05-24 11:04:40    ads_facebook     web              mac   \n",
       "\n",
       "   test  price  converted          city country    lat   long  \n",
       "0     0     39          0       Buffalo     USA  42.89 -78.86  \n",
       "1     0     39          0     Lakeville     USA  44.68 -93.24  \n",
       "2     0     39          0         Parma     USA  41.38 -81.73  \n",
       "3     1     59          0  Fayetteville     USA  35.07 -78.90  \n",
       "4     0     39          0       Fishers     USA  39.95 -86.02  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# merge test_results with user_table\n",
    "data = pd.merge(left=test_results, right=user_table, how='left', on='user_id')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <th>price</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">0</th>\n",
       "      <th>39</th>\n",
       "      <td>202517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">1</th>\n",
       "      <th>39</th>\n",
       "      <td>155</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>113918</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            user_id\n",
       "test price         \n",
       "0    39      202517\n",
       "     59         210\n",
       "1    39         155\n",
       "     59      113918"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Problem with the data: test and price not match\n",
    "data[['price', 'test', 'user_id']].groupby(['test', 'price']).count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It seems that there are some problems with the data\n",
    "* there are a small amount of users who are labelled with `test=0` but actually see price $\\$59$\n",
    "* there are a small amount of users who are labelled with `test=1` but actually see price $\\$39$. \n",
    "\n",
    "Fortunately, they are only a very small portion of the whole dataset. The first step is to remove these problematic inputs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# only keep the valid data\n",
    "data = data[((data['test']==0) & (data['price']==39)) | ((data['test']==1) & (data['price']==59))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create new column called `revenue`\n",
    "data['revenue'] = data['converted'] * data['price']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>user_id</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>source</th>\n",
       "      <th>device</th>\n",
       "      <th>operative_system</th>\n",
       "      <th>test</th>\n",
       "      <th>price</th>\n",
       "      <th>converted</th>\n",
       "      <th>city</th>\n",
       "      <th>country</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>revenue</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>604839</td>\n",
       "      <td>2015-05-08 03:38:34</td>\n",
       "      <td>ads_facebook</td>\n",
       "      <td>mobile</td>\n",
       "      <td>iOS</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>Buffalo</td>\n",
       "      <td>USA</td>\n",
       "      <td>42.89</td>\n",
       "      <td>-78.86</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>624057</td>\n",
       "      <td>2015-05-10 21:08:46</td>\n",
       "      <td>seo-google</td>\n",
       "      <td>mobile</td>\n",
       "      <td>android</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>Lakeville</td>\n",
       "      <td>USA</td>\n",
       "      <td>44.68</td>\n",
       "      <td>-93.24</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>317970</td>\n",
       "      <td>2015-04-04 15:01:23</td>\n",
       "      <td>ads-bing</td>\n",
       "      <td>mobile</td>\n",
       "      <td>android</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>Parma</td>\n",
       "      <td>USA</td>\n",
       "      <td>41.38</td>\n",
       "      <td>-81.73</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>685636</td>\n",
       "      <td>2015-05-07 07:26:01</td>\n",
       "      <td>direct_traffic</td>\n",
       "      <td>mobile</td>\n",
       "      <td>iOS</td>\n",
       "      <td>1</td>\n",
       "      <td>59</td>\n",
       "      <td>0</td>\n",
       "      <td>Fayetteville</td>\n",
       "      <td>USA</td>\n",
       "      <td>35.07</td>\n",
       "      <td>-78.90</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>820854</td>\n",
       "      <td>2015-05-24 11:04:40</td>\n",
       "      <td>ads_facebook</td>\n",
       "      <td>web</td>\n",
       "      <td>mac</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>Fishers</td>\n",
       "      <td>USA</td>\n",
       "      <td>39.95</td>\n",
       "      <td>-86.02</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id            timestamp          source  device operative_system  \\\n",
       "0   604839  2015-05-08 03:38:34    ads_facebook  mobile              iOS   \n",
       "1   624057  2015-05-10 21:08:46      seo-google  mobile          android   \n",
       "2   317970  2015-04-04 15:01:23        ads-bing  mobile          android   \n",
       "3   685636  2015-05-07 07:26:01  direct_traffic  mobile              iOS   \n",
       "4   820854  2015-05-24 11:04:40    ads_facebook     web              mac   \n",
       "\n",
       "   test  price  converted          city country    lat   long  revenue  \n",
       "0     0     39          0       Buffalo     USA  42.89 -78.86        0  \n",
       "1     0     39          0     Lakeville     USA  44.68 -93.24        0  \n",
       "2     0     39          0         Parma     USA  41.38 -81.73        0  \n",
       "3     1     59          0  Fayetteville     USA  35.07 -78.90        0  \n",
       "4     0     39          0       Fishers     USA  39.95 -86.02        0  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user_id                 0\n",
       "timestamp               0\n",
       "source                  0\n",
       "device                  0\n",
       "operative_system        0\n",
       "test                    0\n",
       "price                   0\n",
       "converted               0\n",
       "city                41141\n",
       "country             41141\n",
       "lat                 41141\n",
       "long                41141\n",
       "revenue                 0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's check whether or not the test are totally randomized through simple visualization."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Source"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>ads-bing</th>\n",
       "      <th>ads-google</th>\n",
       "      <th>ads-yahoo</th>\n",
       "      <th>ads_facebook</th>\n",
       "      <th>ads_other</th>\n",
       "      <th>direct_traffic</th>\n",
       "      <th>friend_referral</th>\n",
       "      <th>seo-bing</th>\n",
       "      <th>seo-google</th>\n",
       "      <th>seo-other</th>\n",
       "      <th>seo-yahoo</th>\n",
       "      <th>seo_facebook</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.072014</td>\n",
       "      <td>0.186873</td>\n",
       "      <td>0.023761</td>\n",
       "      <td>0.168253</td>\n",
       "      <td>0.095083</td>\n",
       "      <td>0.191154</td>\n",
       "      <td>0.064790</td>\n",
       "      <td>0.006562</td>\n",
       "      <td>0.074502</td>\n",
       "      <td>0.029030</td>\n",
       "      <td>0.021509</td>\n",
       "      <td>0.066468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.072394</td>\n",
       "      <td>0.188706</td>\n",
       "      <td>0.024219</td>\n",
       "      <td>0.169253</td>\n",
       "      <td>0.092786</td>\n",
       "      <td>0.189259</td>\n",
       "      <td>0.066337</td>\n",
       "      <td>0.007154</td>\n",
       "      <td>0.070753</td>\n",
       "      <td>0.029539</td>\n",
       "      <td>0.021814</td>\n",
       "      <td>0.067786</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      ads-bing  ads-google  ads-yahoo  ads_facebook  ads_other  \\\n",
       "test                                                             \n",
       "0     0.072014    0.186873   0.023761      0.168253   0.095083   \n",
       "1     0.072394    0.188706   0.024219      0.169253   0.092786   \n",
       "\n",
       "      direct_traffic  friend_referral  seo-bing  seo-google  seo-other  \\\n",
       "test                                                                     \n",
       "0           0.191154         0.064790  0.006562    0.074502   0.029030   \n",
       "1           0.189259         0.066337  0.007154    0.070753   0.029539   \n",
       "\n",
       "      seo-yahoo  seo_facebook  \n",
       "test                           \n",
       "0      0.021509      0.066468  \n",
       "1      0.021814      0.067786  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check randominess\n",
    "tmp = data.groupby('test')['source'].apply(lambda x: x.value_counts(normalize=True)).unstack()\n",
    "tmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "tmp.plot(kind='bar', ax=ax)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of source\n",
    "count = data[['source', 'test', 'revenue']].groupby(['source', 'test']).count().reset_index()\n",
    "count.loc[count['test'] == 1, 'revenue'] = count.loc[count['test'] == 1, 'revenue'] / 113918\n",
    "count.loc[count['test'] == 0, 'revenue'] = count.loc[count['test'] == 0, 'revenue'] / 202517\n",
    "\n",
    "fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(18, 6))\n",
    "sns.barplot(x='source', y='revenue', hue='test', data=count, ax=ax[0])\n",
    "ax[0].set_title('Percentage Plot of Source', fontsize=16)\n",
    "\n",
    "sns.barplot(x='source', y='revenue', hue='test', data=data, ax=ax[1])\n",
    "ax[1].set_title('Mean Revenue vs. Source', fontsize=16)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mobile</th>\n",
       "      <th>web</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.594365</td>\n",
       "      <td>0.405635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.578469</td>\n",
       "      <td>0.421531</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        mobile       web\n",
       "test                    \n",
       "0     0.594365  0.405635\n",
       "1     0.578469  0.421531"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check randominess\n",
    "tmp = data.groupby('test')['device'].apply(lambda x: x.value_counts(normalize=True)).unstack()\n",
    "tmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "tmp.plot(kind='bar', ax=ax)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of device\n",
    "count = data[['device', 'test', 'revenue']].groupby(['device', 'test']).count().reset_index()\n",
    "count.loc[count['test'] == 1, 'revenue'] = count.loc[count['test'] == 1, 'revenue'] / 113918\n",
    "count.loc[count['test'] == 0, 'revenue'] = count.loc[count['test'] == 0, 'revenue'] / 202517\n",
    "\n",
    "fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(18, 6))\n",
    "sns.barplot(x='device', y='revenue', hue='test', data=count, ax=ax[0])\n",
    "ax[0].set_title('Percentage Plot of Device', fontsize=16)\n",
    "\n",
    "sns.barplot(x='device', y='revenue', hue='test', data=data, ax=ax[1])\n",
    "ax[1].set_title('Mean Revenue vs. Device', fontsize=16)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Operative System"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>windows</th>\n",
       "      <th>iOS</th>\n",
       "      <th>android</th>\n",
       "      <th>mac</th>\n",
       "      <th>other</th>\n",
       "      <th>linux</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.313638</td>\n",
       "      <td>0.305125</td>\n",
       "      <td>0.237891</td>\n",
       "      <td>0.080709</td>\n",
       "      <td>0.051754</td>\n",
       "      <td>0.010883</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.327718</td>\n",
       "      <td>0.294598</td>\n",
       "      <td>0.234318</td>\n",
       "      <td>0.076459</td>\n",
       "      <td>0.050001</td>\n",
       "      <td>0.016907</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       windows       iOS   android       mac     other     linux\n",
       "test                                                            \n",
       "0     0.313638  0.305125  0.237891  0.080709  0.051754  0.010883\n",
       "1     0.327718  0.294598  0.234318  0.076459  0.050001  0.016907"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check randominess\n",
    "tmp = data.groupby('test')['operative_system'].apply(lambda x: x.value_counts(normalize=True)).unstack()\n",
    "tmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "tmp.plot(kind='bar', ax=ax)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of operative system\n",
    "count = data[['operative_system', 'test', 'revenue']].groupby(['operative_system', 'test']).count().reset_index()\n",
    "count.loc[count['test'] == 1, 'revenue'] = count.loc[count['test'] == 1, 'revenue'] / 113918\n",
    "count.loc[count['test'] == 0, 'revenue'] = count.loc[count['test'] == 0, 'revenue'] / 202517\n",
    "\n",
    "fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(18, 6))\n",
    "sns.barplot(x='operative_system', y='revenue', hue='test', data=count, ax=ax[0])\n",
    "ax[0].set_title('Percentage Plot of Operative System', fontsize=16)\n",
    "\n",
    "sns.barplot(x='operative_system', y='revenue', hue='test', data=data, ax=ax[1])\n",
    "ax[1].set_title('Mean Revenue vs. Operative System', fontsize=16)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From above visualizations, we can see that the test is not $100%$ randomized, especially `device` and `operative_system`. These factors will influence the robostness of our final conclusions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Question 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>converted</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.019900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.015555</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      converted\n",
       "test           \n",
       "0      0.019900\n",
       "1      0.015555"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# check the conversion rate for A/B test\n",
    "data[['test', 'converted']].groupby('test').mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>revenue</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.776083</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.917748</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       revenue\n",
       "test          \n",
       "0     0.776083\n",
       "1     0.917748"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# check the revenue for A/B test\n",
    "data[['test', 'revenue']].groupby('test').mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Clearly, for those with lower price $\\$39$, they have a higher conversion rate $1.990\\%$. For those with higher price $\\$59$, the conversion rate is $1.556\\%$. However, considering the price, the average revenue for higher price ($0.918$) is actually larger than the lower price ($0.776$)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's conduct hypothesis testing.\n",
    "* Null hypothesis H0: experimental group's average revenue is equal to or less than control group's average revenue\n",
    "* Alternative hypothesis H1: experimental group's average revenue is higher than control group's average revenue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ttest_indResult(statistic=-5.715224666463108, pvalue=1.0972577312420781e-08)\n"
     ]
    }
   ],
   "source": [
    "ctr_val = data[data['test'] == 0]['revenue'].values\n",
    "exp_val = data[data['test'] == 1]['revenue'].values\n",
    "print(ttest_ind(ctr_val, exp_val, axis=0, equal_var=False))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The above test is actually two tailed test. What we want is one-tailed test. However, from above result, we can easily conclude that even for one-tailed test, we can safely conclude that we can reject the null hypothesis, and conclude that experimental group's average revenue is higher than control group's average revenue."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Although our test suggest that $\\$59$ price can give us higher average revenue. However, due to the problem with the A/B test, the above result is not scientifically accurate. More scientific A/B test is needed."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Question 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Source"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>source</th>\n",
       "      <th>converted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ads-bing</td>\n",
       "      <td>0.012001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ads-google</td>\n",
       "      <td>0.021502</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ads-yahoo</td>\n",
       "      <td>0.014793</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ads_facebook</td>\n",
       "      <td>0.021198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ads_other</td>\n",
       "      <td>0.014350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>direct_traffic</td>\n",
       "      <td>0.012294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>friend_referral</td>\n",
       "      <td>0.038688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>seo-bing</td>\n",
       "      <td>0.023787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>seo-google</td>\n",
       "      <td>0.016978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>seo-other</td>\n",
       "      <td>0.015686</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>seo-yahoo</td>\n",
       "      <td>0.016226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>seo_facebook</td>\n",
       "      <td>0.016051</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             source  converted\n",
       "0          ads-bing   0.012001\n",
       "1        ads-google   0.021502\n",
       "2         ads-yahoo   0.014793\n",
       "3      ads_facebook   0.021198\n",
       "4         ads_other   0.014350\n",
       "5    direct_traffic   0.012294\n",
       "6   friend_referral   0.038688\n",
       "7          seo-bing   0.023787\n",
       "8        seo-google   0.016978\n",
       "9         seo-other   0.015686\n",
       "10        seo-yahoo   0.016226\n",
       "11     seo_facebook   0.016051"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.groupby('source')['converted'].mean().reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of source\n",
    "fig, ax = plt.subplots(figsize=(16, 6))\n",
    "sns.barplot(x='source', y='converted', data=data, ax=ax)\n",
    "ax.set_title('Mean Conversion Rate vs. Source', fontsize=16)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of source\n",
    "fig, ax = plt.subplots(figsize=(16, 6))\n",
    "sns.barplot(x='source', y='converted', hue='test', data=data, ax=ax)\n",
    "ax.set_title('Mean Conversion Rate vs. Source', fontsize=16)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>device</th>\n",
       "      <th>converted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>mobile</td>\n",
       "      <td>0.018554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>web</td>\n",
       "      <td>0.018023</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   device  converted\n",
       "0  mobile   0.018554\n",
       "1     web   0.018023"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.groupby('device')['converted'].mean().reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of device\n",
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "sns.barplot(x='device', y='converted', data=data, ax=ax)\n",
    "ax.set_title('Mean Conversion Rate vs. Device', fontsize=16)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of device\n",
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "sns.barplot(x='device', y='converted', hue='test', data=data, ax=ax)\n",
    "ax.set_title('Mean Conversion Rate vs. Device', fontsize=16)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Operative System"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>operative_system</th>\n",
       "      <th>converted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>android</td>\n",
       "      <td>0.014919</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>iOS</td>\n",
       "      <td>0.022328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>linux</td>\n",
       "      <td>0.008232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>mac</td>\n",
       "      <td>0.023987</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>other</td>\n",
       "      <td>0.012981</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>windows</td>\n",
       "      <td>0.016966</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  operative_system  converted\n",
       "0          android   0.014919\n",
       "1              iOS   0.022328\n",
       "2            linux   0.008232\n",
       "3              mac   0.023987\n",
       "4            other   0.012981\n",
       "5          windows   0.016966"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.groupby('operative_system')['converted'].mean().reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of operative system\n",
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "sns.barplot(x='operative_system', y='converted', data=data, ax=ax)\n",
    "ax.set_title('Mean Conversion Rate vs. Operative System', fontsize=16)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualization of operative system\n",
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "sns.barplot(x='operative_system', y='converted', hue='test', data=data, ax=ax)\n",
    "ax.set_title('Mean Conversion Rate vs. Operative System', fontsize=16)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From above visualizations, we can easily conclude that price dramatically influence the conversion rate. But beside price, there are still other factors that can influence the conversion rate.\n",
    "* Source\n",
    "    - Friend-referral users have much higher conversion rate\n",
    "* Device\n",
    "    - The influence of device is not significant\n",
    "* Operative System\n",
    "    - Mac and ios users have much higher conversion rate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Actionable Insights (refer to online resource)\n",
    "friend_referral, Apple user (MAC or iOS), low price are three great positive factors which improve the conversion rate\n",
    "if we want to increase the price, which impacts negatively on conversion rate, we must compensate on the other two factors\n",
    "* lauch special marketing program targeted to Apple users (MAC or iOS users)\n",
    "* run some program which rewards user which can invite his/her friend to use our software\n",
    "\n",
    "Linux users don't like our software as much as users on other OS. Development team should find out the reason. For example, is there any incompatibility issue on Linux?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}