final_project_troitskiy_var3_071122.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Структура\n",
    "* 1. [Подготовка данных (Предварительный анализ)](#podgotovka)\n",
    "* 2. [Проверка системы сплитования](#proverka)\n",
    "* 3. [Анализ результатов ААБ Теста](#analis)\n",
    "    * 3.1 Общие показатели (Был ли эксперимент успешен в целом)\n",
    "    * 3.2 [Анализ характеристик групп пользователей](#groups)\n",
    "        * 3.2.1 [Гендерный состав](#gender)\n",
    "        * 3.2.2 [Привлекательность](#attraction)\n",
    "        * 3.2.3 [Количество посещений](#visits)\n",
    "        * 3.2.4 [Страны](#country)\n",
    "        * 3.2.5 [Возраст](#ages)\n",
    "* 4. [Выводы](#vivod) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Подготовка данных (Предварительный анализ)<a name=\"podgotovka\"></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "# from datetime import timedelta\n",
    "#from datetime import datetime\n",
    "# from io import StringIO"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns\n",
    "from scipy import stats\n",
    "from scipy.stats import levene\n",
    "# from scipy.stats import wilcoxon\n",
    "import matplotlib.pyplot as plt\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [],
   "source": [
    "import statsmodels.formula.api as smf\n",
    "from statsmodels.stats.api import anova_lm\n",
    "from statsmodels.formula.api import ols\n",
    "from statsmodels.stats.multicomp import (pairwise_tukeyhsd,\n",
    "                                         MultiComparison)\n",
    "import scikit_posthocs as sp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from sklearn.linear_model import LinearRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm.auto import tqdm\n",
    "from scipy.stats import norm\n",
    "plt.style.use('ggplot')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_bootstrap(\n",
    "    data_column_1,  # числовые значения первой выборки\n",
    "    data_column_2,  # числовые значения второй выборки\n",
    "    boot_it=1000,  # количество бутстрэп-подвыборок\n",
    "    statistic=np.mean,  # интересующая нас статистика\n",
    "    bootstrap_conf_level=0.95  # уровень значимости\n",
    "): \n",
    "    boot_data = []\n",
    "    for i in tqdm(range(boot_it)):  # извлекаем подвыборки\n",
    "        samples_1 = data_column_1.sample(\n",
    "            len(data_column_1), \n",
    "            replace=True  # параметр возвращения\n",
    "        ).values\n",
    "        \n",
    "        samples_2 = data_column_2.sample(\n",
    "            len(data_column_1), \n",
    "            replace=True\n",
    "        ).values\n",
    "        \n",
    "        boot_data.append(statistic(samples_1) - statistic(samples_2))  # mean() - применяем статистику\n",
    "        \n",
    "    pd_boot_data = pd.DataFrame(boot_data)\n",
    "        \n",
    "    left_quant = (1 - bootstrap_conf_level) / 2\n",
    "    right_quant = 1 - (1 - bootstrap_conf_level) / 2\n",
    "    quants = pd_boot_data.quantile([left_quant, right_quant])\n",
    "        \n",
    "    p_1 = norm.cdf(\n",
    "        x=0, \n",
    "        loc=np.mean(boot_data), \n",
    "        scale=np.std(boot_data)\n",
    "    )\n",
    "    p_2 = norm.cdf(\n",
    "        x=0, \n",
    "        loc =-np.mean(boot_data), \n",
    "        scale=np.std(boot_data)\n",
    "    )\n",
    "    p_value = min(p_1, p_2) * 2\n",
    "        \n",
    "    # Визуализация\n",
    "    _, _, bars = plt.hist(pd_boot_data[0], bins=50)\n",
    "    for bar in bars: \n",
    "        if bar.get_x() <= quants.iloc[0][0] or bar.get_x() >= quants.iloc[1][0]: \n",
    "            bar.set_facecolor('red')\n",
    "        else: \n",
    "            bar.set_facecolor('grey')\n",
    "            bar.set_edgecolor('black')\n",
    "    \n",
    "    plt.style.use('ggplot')\n",
    "    plt.vlines(quants, ymin=0, ymax=50, linestyle='--')\n",
    "    plt.xlabel('boot_data')\n",
    "    plt.ylabel('frequency')\n",
    "    plt.title(\"Histogram of boot_data\")\n",
    "    plt.show()\n",
    "       \n",
    "    return {  # \"boot_data\": boot_data, \n",
    "            \"quants\": quants, \n",
    "            \"p_value\": p_value}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "from urllib.parse import urlencode"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_df_by_yandex_link(link, name): \n",
    "    base_url = 'https: //cloud-api.yandex.net/v1/disk/public/resources/download?'\n",
    "    public_key = link  # This inscribe your link\n",
    "    # Get the download link\n",
    "    final_url = base_url + urlencode(dict(public_key=public_key))\n",
    "    response = requests.get (final_url)\n",
    "    download_url = response.json()['href']\n",
    "    # Load the file and save it\n",
    "    download_response = requests.get(download_url)\n",
    "    with open(name, 'wb') as f:  # Here specify the correct path to the file\n",
    "        f.write(download_response.content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "yandex_links = ['https: //disk.yandex.ru/d/4XXIME4osGrMRA', 'https: //disk.yandex.ru/d/gvCWpZ55ODzs2g'\n",
    "                'https: //disk.yandex.ru/d/yJFydMNNGkEKfg', 'https: //disk.yandex.ru/d/VY5W0keMX5TZBQ'\n",
    "                'https: //disk.yandex.ru/d/br6KkQupzzTGoQ', 'https: //disk.yandex.ru/d/th5GL0mGOc-qzg']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "names = ['users_test', 'transactions_test', 'users_control_1.csv', 'transactions_control_1.csv',\n",
    "         'users_control_2.csv', 'transactions_control_2.csv']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "h = 0\n",
    "for i in yandex_links: \n",
    "    name = names[i]\n",
    "    get_df_by_yandex_link(i, name)\n",
    "    h += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "users_test = pd.read_csv('users_test.csv', sep=';')\n",
    "transactions_test = pd.read_csv('transactions_test.csv', sep=';')\n",
    "\n",
    "users_control_1 = pd.read_csv('users_control_1.csv', sep=';')\n",
    "transactions_control_1 = pd.read_csv('transactions_control_1.csv', sep=';')\n",
    "\n",
    "users_control_2 = pd.read_csv('users_control_2.csv', sep=';')\n",
    "transactions_control_2 = pd.read_csv('transactions_control_2.csv', sep=';')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "uid             630\n",
       "country         630\n",
       "joined_at         0\n",
       "paid_at           0\n",
       "revenue         630\n",
       "payment_id      630\n",
       "from_page       630\n",
       "product_type    630\n",
       "dtype: int64"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "transactions_control_1.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "transactions_control_1 = transactions_control_1.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "nunique_comparison = pd.concat([users_test.nunique().to_frame(),\n",
    "                                users_control_1.nunique().to_frame(),\n",
    "                                users_control_2.nunique().to_frame()],\n",
    "                                axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "nunique_comparison.columns = ['test', 'control_1', 'control_2']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "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>test</th>\n",
       "      <th>control_1</th>\n",
       "      <th>control_2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>uid</th>\n",
       "      <td>4308</td>\n",
       "      <td>4340</td>\n",
       "      <td>4264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>country</th>\n",
       "      <td>51</td>\n",
       "      <td>40</td>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age</th>\n",
       "      <td>57</td>\n",
       "      <td>56</td>\n",
       "      <td>59</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>views_count</th>\n",
       "      <td>331</td>\n",
       "      <td>339</td>\n",
       "      <td>328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>attraction_coeff</th>\n",
       "      <td>600</td>\n",
       "      <td>587</td>\n",
       "      <td>585</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  test  control_1  control_2\n",
       "uid               4308       4340       4264\n",
       "country             51         40         45\n",
       "age                 57         56         59\n",
       "views_count        331        339        328\n",
       "attraction_coeff   600        587        585"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nunique_comparison.loc[['uid', 'country', 'age', 'views_count', 'attraction_coeff']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Для начала необходимо убедиться в корректной настройке системы сплитования. Выше в таблице представлено сравнение уникальных значений для id, стран, возраста, параметров привлекательности пользователей. В связи с тем, что во второй контрольной группе представлено на 5 стран больше, чем в первой можно сделать предположение о целосообразности попарного сравнения результатов вместо однофакторного дисперсионного анализа. Однако необходимо уточнить состав стран и количество пользователей, соответствующее каждой стране**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Проверка системы сплитования<a name=\"proverka\"></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_comparison = pd.concat([users_control_2.groupby('country').agg({'uid': 'count'})\n",
    "                    .sort_values('uid', ascending=False), users_control_1.groupby('country').agg({'uid': 'count'})\n",
    "                    .sort_values('uid', ascending=False)], \\\n",
    "                    axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_comparison.columns = ['control_2', 'control_1']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "control_2    24.0\n",
       "control_1    19.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "country_comparison.query(\"control_2=='NaN' | control_1=='NaN'\").sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**В первой контрольной группе присутствует 19 пользователей из стран, которые не представлены во второй группе. И 24 пользователя соответственно во второй.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "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>control_2</th>\n",
       "      <th>control_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>United States of America</th>\n",
       "      <td>1486.0</td>\n",
       "      <td>1471.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Argentina</th>\n",
       "      <td>397.0</td>\n",
       "      <td>389.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France</th>\n",
       "      <td>376.0</td>\n",
       "      <td>370.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Spain</th>\n",
       "      <td>310.0</td>\n",
       "      <td>342.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Italy</th>\n",
       "      <td>293.0</td>\n",
       "      <td>338.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Chile</th>\n",
       "      <td>286.0</td>\n",
       "      <td>287.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>United Arab Emirates</th>\n",
       "      <td>211.0</td>\n",
       "      <td>207.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Germany</th>\n",
       "      <td>202.0</td>\n",
       "      <td>188.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>United Kingdom (Great Britain)</th>\n",
       "      <td>146.0</td>\n",
       "      <td>146.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Canada</th>\n",
       "      <td>134.0</td>\n",
       "      <td>134.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Portugal</th>\n",
       "      <td>100.0</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Belgium</th>\n",
       "      <td>90.0</td>\n",
       "      <td>108.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Israel</th>\n",
       "      <td>78.0</td>\n",
       "      <td>109.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Australia</th>\n",
       "      <td>51.0</td>\n",
       "      <td>50.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Switzerland</th>\n",
       "      <td>35.0</td>\n",
       "      <td>42.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                control_2  control_1\n",
       "United States of America           1486.0     1471.0\n",
       "Argentina                           397.0      389.0\n",
       "France                              376.0      370.0\n",
       "Spain                               310.0      342.0\n",
       "Italy                               293.0      338.0\n",
       "Chile                               286.0      287.0\n",
       "United Arab Emirates                211.0      207.0\n",
       "Germany                             202.0      188.0\n",
       "United Kingdom (Great Britain)      146.0      146.0\n",
       "Canada                              134.0      134.0\n",
       "Portugal                            100.0      100.0\n",
       "Belgium                              90.0      108.0\n",
       "Israel                               78.0      109.0\n",
       "Australia                            51.0       50.0\n",
       "Switzerland                          35.0       42.0"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "country_comparison.head(15)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Удельная доля пользователей из уникальных для каждой контрольной группы стран позволяет проводить дальнейший анализ данных.** "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Дополнительно сравним различия в страницах откуда пользователь перешел на страницу оплаты.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "from_page_comparison = pd.concat([transactions_test.groupby('from_page').agg({'payment_id': 'count'}),\n",
    "                                  transactions_control_1.groupby('from_page').agg({'payment_id': 'count'}),\n",
    "                                  transactions_control_2.groupby('from_page').agg({'payment_id': 'count'})],\n",
    "                                  axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "from_page_comparison.columns = ['test', 'control_1', 'control_2']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>test</th>\n",
       "      <th>control_1</th>\n",
       "      <th>control_2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>coins_popup</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>delivery_vip</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>empty_mutual</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>likes</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>super_popup</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>admins</th>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>empty_admins</th>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>no_sympathy</th>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>refund_VP</th>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>refund_WP</th>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>refung_AP</th>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              test  control_1  control_2\n",
       "coins_popup    1.0        2.0        NaN\n",
       "delivery_vip   1.0        NaN        NaN\n",
       "empty_mutual   1.0        NaN        1.0\n",
       "likes          1.0        NaN        NaN\n",
       "super_popup    2.0        2.0        NaN\n",
       "admins         NaN        2.0        NaN\n",
       "empty_admins   NaN        5.0        NaN\n",
       "no_sympathy    NaN        2.0        NaN\n",
       "refund_VP      NaN        1.0        NaN\n",
       "refund_WP      NaN        1.0        NaN\n",
       "refung_AP      NaN        1.0        NaN"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from_page_comparison.query(\"control_2=='NaN' | control_1=='NaN' | test=='NaN'\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "from_page_comparison['average'] = from_page_comparison.test + from_page_comparison.control_1 \\\n",
    "                                                            + from_page_comparison.control_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "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>test</th>\n",
       "      <th>control_1</th>\n",
       "      <th>control_2</th>\n",
       "      <th>average</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>empty_likes</th>\n",
       "      <td>72.0</td>\n",
       "      <td>99.0</td>\n",
       "      <td>101.0</td>\n",
       "      <td>272.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>trial_vip_popup</th>\n",
       "      <td>48.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>149.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>none</th>\n",
       "      <td>34.0</td>\n",
       "      <td>34.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>autoreply_message</th>\n",
       "      <td>27.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>80.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>menu</th>\n",
       "      <td>27.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>72.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>promo_09</th>\n",
       "      <td>18.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>61.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>autorefill</th>\n",
       "      <td>5.0</td>\n",
       "      <td>23.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>48.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   test  control_1  control_2  average\n",
       "empty_likes        72.0       99.0      101.0    272.0\n",
       "trial_vip_popup    48.0       56.0       45.0    149.0\n",
       "none               34.0       34.0       32.0    100.0\n",
       "autoreply_message  27.0       29.0       24.0     80.0\n",
       "menu               27.0       31.0       14.0     72.0\n",
       "promo_09           18.0       21.0       22.0     61.0\n",
       "autorefill          5.0       23.0       20.0     48.0"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from_page_comparison.sort_values('average', ascending=False).head(7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Количество пользователей, перешедших с уникальных для каждой группы страниц, существенно не влияет  на однородность составов групп.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_comparison = pd.concat([users_test.groupby('gender').agg({'uid': 'count'}),\n",
    "                            users_control_1.groupby('gender').agg({'uid': 'count'}),\n",
    "                            users_control_2.groupby('gender').agg({'uid': 'count'})],\n",
    "                            axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_comparison.columns = ['test', 'control_1', 'control_2']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_comparison = age_comparison.T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_df = pd.DataFrame({'test': (age_comparison.loc['test'][0] / age_comparison.loc['test'].sum()).round(3),\n",
    "          'control_1': (age_comparison.loc['control_1'][0] / age_comparison.loc['control_1'].sum()).round(3),\n",
    "          'control_2': (age_comparison.loc['control_2'][0] / age_comparison.loc['control_2'].sum()).round(3)},\n",
    "          index=['age_relation'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "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>test</th>\n",
       "      <th>control_1</th>\n",
       "      <th>control_2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>age_relation</th>\n",
       "      <td>0.047</td>\n",
       "      <td>0.042</td>\n",
       "      <td>0.046</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               test  control_1  control_2\n",
       "age_relation  0.047      0.042      0.046"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "age_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Отношение количества пользователей определенного пола к другому примерно одинаковое во всех трех группах.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "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>test</th>\n",
       "      <th>control_1</th>\n",
       "      <th>control_2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>age_filter_start_mean</th>\n",
       "      <td>26.701</td>\n",
       "      <td>26.794</td>\n",
       "      <td>26.954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age_filter_end_mean</th>\n",
       "      <td>38.249</td>\n",
       "      <td>38.553</td>\n",
       "      <td>38.907</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         test  control_1  control_2\n",
       "age_filter_start_mean  26.701     26.794     26.954\n",
       "age_filter_end_mean    38.249     38.553     38.907"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame({'test': (users_test.age_filter_start.mean().round(3), users_test.age_filter_end.mean().round(3)),\n",
    "     'control_1': (users_control_1.age_filter_start.mean().round(3), users_control_1.age_filter_end.mean().round(3)),\n",
    "    'control_2': (users_control_2.age_filter_start.mean().round(3), users_control_2.age_filter_end.mean().round(3))},\n",
    "            index=['age_filter_start_mean', 'age_filter_end_mean'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Средний возраст в фильтрах поиска также совпадает во всех случаях**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KstestResult(statistic=0.1674063650140723, pvalue=0.038402920239277494)"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.kstest(list(transactions_control_1[transactions_control_1.product_type=='coins'].revenue), \\\n",
    "                  list(transactions_control_2[transactions_control_2.product_type=='coins'].revenue))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Дополнительная проверка системы сплитования в части распределений выручки за внутреннюю валюту для последующего пофакторного анализа (пункт 3.1).\n",
    "Cтатистика теста равна 0.167, а соответствующее значение p равно .038.Поскольку p-значение меньше 0,05, мы отвергаем нулевую гипотезу. У нас есть достаточно доказательств, чтобы сказать, что два выборочных набора из контрольных групп не происходят из одного и того же распределения.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "aa1 = np.array([[users_control_2.is_premium.sum(), users_control_2.shape[0] - users_control_2.is_premium.sum()],\n",
    "                [users_control_1.is_premium.sum(), users_control_1.shape[0] - users_control_1.is_premium.sum()]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9008571305798496\n"
     ]
    }
   ],
   "source": [
    "print(stats.chi2_contingency(aa1, correction=False)[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Тестирование данных АА теста показывает отсутствие значимых различий в контрольных группах**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "control_full = pd.concat([users_control_1, users_control_2],\n",
    "                          axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "control_transactions_full = pd.concat([transactions_control_1, transactions_control_2], axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(705, 273)"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "control_transactions_full.shape[0], transactions_test.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "control_full['group'] = 'control'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "users_test['group'] = 'test'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_data = pd.concat([control_full, users_test],\n",
    "                       axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Анализ результатов ААБ Теста<a name=\"analis\"></a>\n",
    "## 3.1 Общие показатели (Был ли эксперимент успешен в целом)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [],
   "source": [
    "general_indicators = pd.DataFrame({'Control': [control_full.was_premium.sum() / control_full.uid.nunique(),\n",
    "                                    control_full.is_premium.sum() / control_full.uid.nunique(),\n",
    "                                    control_full.total_revenue.sum() / control_full.uid.nunique()],\n",
    "                                    'Test': [users_test.was_premium.sum() / users_test.uid.nunique(),\n",
    "                                    users_test.is_premium.sum() / users_test.uid.nunique(),\n",
    "                                    users_test.total_revenue.sum() / users_test.uid.nunique()]},\n",
    "                                    index=['ex_premiums_on_uid', 'premiums_on_uid', 'revenue_on_uid'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "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>Control</th>\n",
       "      <th>Test</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ex_premiums_on_uid</th>\n",
       "      <td>0.098443</td>\n",
       "      <td>0.094708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>premiums_on_uid</th>\n",
       "      <td>0.044514</td>\n",
       "      <td>0.036444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>revenue_on_uid</th>\n",
       "      <td>523.210716</td>\n",
       "      <td>534.080316</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       Control        Test\n",
       "ex_premiums_on_uid    0.098443    0.094708\n",
       "premiums_on_uid       0.044514    0.036444\n",
       "revenue_on_uid      523.210716  534.080316"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "general_indicators"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Предварительный расчет таких показателей как количество клиентов с бывшей и действующей подпиской на пользователя и выручки на количество пользователей (ARPU).**"
   ]
  },
  {
   "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>control_full</th>\n",
       "      <th>test</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>clients</th>\n",
       "      <td>379</td>\n",
       "      <td>146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>users</th>\n",
       "      <td>8604</td>\n",
       "      <td>4308</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         control_full  test\n",
       "clients           379   146\n",
       "users            8604  4308"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame({'control_full': (control_full[control_full.total_revenue  !=  0].shape[0], control_full.shape[0]),\n",
    "              'test': (users_test[users_test.total_revenue  !=  0].shape[0], users_test.shape[0])},\n",
    "              index=['clients', 'users'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Относительно небольшой процент клиентов из пользователей совершили покупки**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LeveneResult(statistic=0.09659654815825328, pvalue=0.7559577996728548)"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "levene(list(users_test.total_revenue),\n",
    "       list(control_full.total_revenue),\n",
    "       center='mean')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Проверим имеют ли выборки данных о выручке равные дисперсии. Р-значение составляет более 0,05. Это означает, что целесообразно не отвергать нулевую гипотезу, требование гомогенности дисперсий удовлетворено для последующего дисперсионного анализа средних нормированной выручки.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "F_onewayResult(statistic=0.011111300162398436, pvalue=0.916051998389317)"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.f_oneway(list(users_test.total_revenue),\n",
    "               list(control_full.total_revenue))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Так как p-значение больше 0.05, то мы не отвергаем нулевую гипотезу и делаем вывод, что средние выручки для групп значимо не отличаются.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ShapiroResult(statistic=0.9119991064071655, pvalue=4.6361325152163846e-14)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.shapiro(np.log(control_full[control_full.total_revenue != 0].total_revenue))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.FacetGrid at 0x7fbb31453ef0>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.displot(np.log(control_full[control_full.total_revenue != 0].total_revenue), kde=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Нулевая гипотеза в случае Шапиро-Уилка — нормальность распределения, распределение выручки для контрольной группы явно отклоняется от нормального (даже без учета нулей), аналогично и в тестовой группе**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6b5d431eadb14ad2b0722ec4f1618adf",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=1000.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "{'quants':                  0\n",
       " 0.025   592.916667\n",
       " 0.975  3809.320238,\n",
       " 'p_value': 0.007911663832772245}"
      ]
     },
     "execution_count": 138,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_bootstrap(\n",
    "    transactions_test.revenue,  # числовые значения первой выборки\n",
    "    control_transactions_full.revenue,  # числовые значения второй выборки\n",
    "    boot_it=1000,  # количество бутстрэп-подвыборок\n",
    "    statistic=np.mean,  # интересующая нас статистика (средние)\n",
    "    bootstrap_conf_level=0.95)  # уровень значимости "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KstestResult(statistic=0.25656612890655445, pvalue=6.844746991419015e-12)"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.ks_2samp(transactions_test.revenue, control_transactions_full.revenue)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Однако согласно данным о транзакциях разница в средних значениях выручки подтверждается вышеуказанными бутстрап-тестом и проверкой по критерию Колмогорова-Смирнова**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "premchi = np.array([[users_test.is_premium.sum(), users_test.shape[0] - users_test.is_premium.sum()],\n",
    "                [control_full.is_premium.sum(), control_full.shape[0] - control_full.is_premium.sum()]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.030770049237882124\n"
     ]
    }
   ],
   "source": [
    "print(stats.chi2_contingency(premchi, correction=False)[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Дополнительно проведем попарное сравнение наличия действующей подписки у пользователей в тестовой и контрольной группах. \n",
    "p-value менее уровня 5% подтверждает статзначимые различия между тестовой и контрольной группами. Таким образом цену за подписку повысили, выручка увеличилась, но количество подписок уменьшилось.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Далее попробуем найти различия в группах пользователей по типу продукта"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "product_type_comparison_avg = pd.concat([control_transactions_full.groupby('product_type')\n",
    "                                                                  .agg({'revenue': lambda x: x.mean().round(2)}),\n",
    "                                     transactions_test.groupby('product_type')\n",
    "                                                                  .agg({'revenue': lambda x: x.mean().round(2)})],\n",
    "                                     axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [],
   "source": [
    "product_type_comparison_avg.columns = ['control', 'test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "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>control</th>\n",
       "      <th>test</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>product_type</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>coins</th>\n",
       "      <td>4669.67</td>\n",
       "      <td>5341.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>other_type</th>\n",
       "      <td>3203.65</td>\n",
       "      <td>5786.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>premium_no_trial</th>\n",
       "      <td>7155.25</td>\n",
       "      <td>10667.18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>trial_premium</th>\n",
       "      <td>11805.84</td>\n",
       "      <td>12906.48</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   control      test\n",
       "product_type                        \n",
       "coins              4669.67   5341.28\n",
       "other_type         3203.65   5786.79\n",
       "premium_no_trial   7155.25  10667.18\n",
       "trial_premium     11805.84  12906.48"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "product_type_comparison_avg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "pt_comp = pd.DataFrame({'control_payments': control_transactions_full.groupby('product_type')\n",
    "                                                                     .agg({'payment_id': 'count'}).payment_id,\n",
    "                        'control_uids':control_transactions_full.groupby('product_type').agg({'uid': 'nunique'}).uid,\n",
    "                        'test_payments': transactions_test.groupby('product_type')\n",
    "                                                          .agg({'payment_id': 'count'}).payment_id,\n",
    "                        'test_uids': transactions_test.groupby('product_type').agg({'uid': 'nunique'}).uid},\n",
    "                        index=['coins', 'other_type', 'premium_no_trial', 'trial_premium'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "pt_comp['control_payments_on_uid'] = (pt_comp.control_payments / pt_comp.control_uids).round(2)\n",
    "pt_comp['test_payments_on_uid'] = (pt_comp.test_payments / pt_comp.test_uids).round(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "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>control_payments</th>\n",
       "      <th>control_uids</th>\n",
       "      <th>test_payments</th>\n",
       "      <th>test_uids</th>\n",
       "      <th>control_payments_on_uid</th>\n",
       "      <th>test_payments_on_uid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>coins</th>\n",
       "      <td>273</td>\n",
       "      <td>134</td>\n",
       "      <td>106</td>\n",
       "      <td>52</td>\n",
       "      <td>2.04</td>\n",
       "      <td>2.04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>other_type</th>\n",
       "      <td>99</td>\n",
       "      <td>63</td>\n",
       "      <td>29</td>\n",
       "      <td>19</td>\n",
       "      <td>1.57</td>\n",
       "      <td>1.53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>premium_no_trial</th>\n",
       "      <td>213</td>\n",
       "      <td>201</td>\n",
       "      <td>76</td>\n",
       "      <td>67</td>\n",
       "      <td>1.06</td>\n",
       "      <td>1.13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>trial_premium</th>\n",
       "      <td>120</td>\n",
       "      <td>120</td>\n",
       "      <td>62</td>\n",
       "      <td>61</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  control_payments  control_uids  test_payments  test_uids  \\\n",
       "coins                          273           134            106         52   \n",
       "other_type                      99            63             29         19   \n",
       "premium_no_trial               213           201             76         67   \n",
       "trial_premium                  120           120             62         61   \n",
       "\n",
       "                  control_payments_on_uid  test_payments_on_uid  \n",
       "coins                                2.04                  2.04  \n",
       "other_type                           1.57                  1.53  \n",
       "premium_no_trial                     1.06                  1.13  \n",
       "trial_premium                        1.00                  1.02  "
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pt_comp"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Сравнение средней выручки по типу продукта показывает увеличение показателей относительно контрольной группы. Количество платежей на пользователя при этом не отличается.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "uid\n",
       "False    4162\n",
       "Name: payment_id, dtype: int64"
      ]
     },
     "execution_count": 173,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "users_test.merge(transactions_test, on=['uid'], how='left').isna().groupby('uid').sum().payment_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.39"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(((users_test.shape[0] - 4162) / users_test.shape[0] * 100), 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "uid\n",
       "False    8224\n",
       "Name: payment_id, dtype: int64"
      ]
     },
     "execution_count": 174,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "control_full.merge(control_transactions_full, on=['uid'], how='left').isna().groupby('uid').sum().payment_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.42"
      ]
     },
     "execution_count": 152,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(((control_full.shape[0] - 8224) / control_full.shape[0] * 100), 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Процент конверсии в покупку уменьшился приблизительно на 1% (доля клиентов от пользователей)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.6443825441039928"
      ]
     },
     "execution_count": 170,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "users_test.query(\"was_premium == 1 & is_premium == 1\").shape[0] / users_test.shape[0] * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.451417945141794"
      ]
     },
     "execution_count": 171,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "control_full.query(\"was_premium == 1 & is_premium == 1\").shape[0] / control_full.shape[0] * 100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Приблизительно на 0,8 уменьшился процент клиентов, не отменивших премиум подписку.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "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>uid</th>\n",
       "      <th>country</th>\n",
       "      <th>joined_at</th>\n",
       "      <th>paid_at</th>\n",
       "      <th>revenue</th>\n",
       "      <th>payment_id</th>\n",
       "      <th>from_page</th>\n",
       "      <th>product_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>891345942</td>\n",
       "      <td>Italy</td>\n",
       "      <td>2017-05-11 13:00:00</td>\n",
       "      <td>2017-11-13 15:04:00</td>\n",
       "      <td>12909</td>\n",
       "      <td>147</td>\n",
       "      <td>trial_vip_popup</td>\n",
       "      <td>trial_premium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>892054251</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>2017-10-22 00:33:00</td>\n",
       "      <td>2017-10-30 01:37:00</td>\n",
       "      <td>13923</td>\n",
       "      <td>147</td>\n",
       "      <td>trial_vip_popup</td>\n",
       "      <td>trial_premium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>892236423</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>2017-10-18 01:09:00</td>\n",
       "      <td>2017-10-23 00:15:00</td>\n",
       "      <td>3783</td>\n",
       "      <td>67</td>\n",
       "      <td>menu</td>\n",
       "      <td>other_type</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>892236423</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>2017-10-18 01:09:00</td>\n",
       "      <td>2017-10-27 22:38:00</td>\n",
       "      <td>3783</td>\n",
       "      <td>67</td>\n",
       "      <td>menu</td>\n",
       "      <td>other_type</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>892168170</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>2017-10-19 17:10:00</td>\n",
       "      <td>2017-10-27 19:10:00</td>\n",
       "      <td>9087</td>\n",
       "      <td>147</td>\n",
       "      <td>trial_vip_popup</td>\n",
       "      <td>trial_premium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>891419625</td>\n",
       "      <td>Israel</td>\n",
       "      <td>2017-04-11 05:58:00</td>\n",
       "      <td>2017-11-11 08:33:00</td>\n",
       "      <td>14352</td>\n",
       "      <td>147</td>\n",
       "      <td>trial_vip_popup</td>\n",
       "      <td>trial_premium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269</th>\n",
       "      <td>891447105</td>\n",
       "      <td>United Kingdom (Great Britain)</td>\n",
       "      <td>2017-03-11 17:05:00</td>\n",
       "      <td>2017-11-11 04:23:00</td>\n",
       "      <td>7319</td>\n",
       "      <td>147</td>\n",
       "      <td>fans</td>\n",
       "      <td>premium_no_trial</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>891497073</td>\n",
       "      <td>United States of America</td>\n",
       "      <td>2017-02-11 15:33:00</td>\n",
       "      <td>2017-05-11 04:06:00</td>\n",
       "      <td>2730</td>\n",
       "      <td>67</td>\n",
       "      <td>empty_likes</td>\n",
       "      <td>other_type</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>891406224</td>\n",
       "      <td>Israel</td>\n",
       "      <td>2017-04-11 12:56:00</td>\n",
       "      <td>2017-04-11 13:35:00</td>\n",
       "      <td>2743</td>\n",
       "      <td>67</td>\n",
       "      <td>empty_likes</td>\n",
       "      <td>other_type</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>272</th>\n",
       "      <td>891499359</td>\n",
       "      <td>Germany</td>\n",
       "      <td>2017-02-11 14:25:00</td>\n",
       "      <td>2017-02-11 14:30:00</td>\n",
       "      <td>1599</td>\n",
       "      <td>19</td>\n",
       "      <td>none</td>\n",
       "      <td>coins</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>273 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           uid                         country            joined_at  \\\n",
       "0    891345942                           Italy  2017-05-11 13:00:00   \n",
       "1    892054251        United States of America  2017-10-22 00:33:00   \n",
       "2    892236423        United States of America  2017-10-18 01:09:00   \n",
       "3    892236423        United States of America  2017-10-18 01:09:00   \n",
       "4    892168170        United States of America  2017-10-19 17:10:00   \n",
       "..         ...                             ...                  ...   \n",
       "268  891419625                          Israel  2017-04-11 05:58:00   \n",
       "269  891447105  United Kingdom (Great Britain)  2017-03-11 17:05:00   \n",
       "270  891497073        United States of America  2017-02-11 15:33:00   \n",
       "271  891406224                          Israel  2017-04-11 12:56:00   \n",
       "272  891499359                         Germany  2017-02-11 14:25:00   \n",
       "\n",
       "                 paid_at  revenue  payment_id        from_page  \\\n",
       "0    2017-11-13 15:04:00    12909         147  trial_vip_popup   \n",
       "1    2017-10-30 01:37:00    13923         147  trial_vip_popup   \n",
       "2    2017-10-23 00:15:00     3783          67             menu   \n",
       "3    2017-10-27 22:38:00     3783          67             menu   \n",
       "4    2017-10-27 19:10:00     9087         147  trial_vip_popup   \n",
       "..                   ...      ...         ...              ...   \n",
       "268  2017-11-11 08:33:00    14352         147  trial_vip_popup   \n",
       "269  2017-11-11 04:23:00     7319         147             fans   \n",
       "270  2017-05-11 04:06:00     2730          67      empty_likes   \n",
       "271  2017-04-11 13:35:00     2743          67      empty_likes   \n",
       "272  2017-02-11 14:30:00     1599          19             none   \n",
       "\n",
       "         product_type  \n",
       "0       trial_premium  \n",
       "1       trial_premium  \n",
       "2          other_type  \n",
       "3          other_type  \n",
       "4       trial_premium  \n",
       "..                ...  \n",
       "268     trial_premium  \n",
       "269  premium_no_trial  \n",
       "270        other_type  \n",
       "271        other_type  \n",
       "272             coins  \n",
       "\n",
       "[273 rows x 8 columns]"
      ]
     },
     "execution_count": 178,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "transactions_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [],
   "source": [
    "transactions_test.joined_at = pd.to_datetime(transactions_test.joined_at,\n",
    "                                             format='%Y-%m-%d %H:%M:%S')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [],
   "source": [
    "transactions_test.paid_at = pd.to_datetime(transactions_test.paid_at,\n",
    "                                             format='%Y-%m-%d %H:%M:%S') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [],
   "source": [
    "transactions_test['join_pay_diff'] = transactions_test.paid_at - transactions_test.joined_at"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Timedelta('2 days 07:32:00')"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "transactions_test.join_pay_diff.median()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [],
   "source": [
    "control_transactions_full.joined_at = pd.to_datetime(control_transactions_full.joined_at,\n",
    "                                                  format='%Y-%m-%d %H:%M:%S')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "control_transactions_full.paid_at = pd.to_datetime(control_transactions_full.paid_at,\n",
    "                                                format='%Y-%m-%d %H:%M:%S') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [],
   "source": [
    "control_transactions_full['join_pay_diff'] = control_transactions_full.paid_at - control_transactions_full.joined_at"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Timedelta('1 days 20:09:00')"
      ]
     },
     "execution_count": 188,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "control_transactions_full.join_pay_diff.median()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Медианная разница времени между регистрацией и покупкой увеличилась примерно на 11 часов.**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3.2 Анализ характеристик групп пользователей<a name=\"groups\"></a>\n",
    "## 3.2.1 Гендерный состав"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_rev_by_gender = pd.concat([control_full.groupby('gender').agg({'total_revenue': lambda x: x.mean().round(2)}),\n",
    "                               users_test.groupby('gender').agg({'total_revenue': lambda x: x.mean().round(2)})],\n",
    "                               axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_rev_by_gender.columns = ['control', 'test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>control</th>\n",
       "      <th>test</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gender</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>228.83</td>\n",
       "      <td>134.05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>536.70</td>\n",
       "      <td>553.76</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        control    test\n",
       "gender                 \n",
       "0        228.83  134.05\n",
       "1        536.70  553.76"
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "age_rev_by_gender"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_revenue_comparison = pd.concat([users_test.groupby('gender').agg({'total_revenue': 'sum'}),\n",
    "                                      control_full.groupby('gender').agg({'total_revenue': 'sum'})],\n",
    "                                      axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [],
   "source": [
    "total_revenue_comparison.columns = ['control', 'test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "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>control</th>\n",
       "      <th>test</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gender</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>27079</td>\n",
       "      <td>86268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2273739</td>\n",
       "      <td>4415437</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        control     test\n",
       "gender                  \n",
       "0         27079    86268\n",
       "1       2273739  4415437"
      ]
     },
     "execution_count": 200,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_revenue_comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "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>test</th>\n",
       "      <th>control</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>clients_gender_1</th>\n",
       "      <td>144</td>\n",
       "      <td>368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>clients_gender_0</th>\n",
       "      <td>2</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  test  control\n",
       "clients_gender_1   144      368\n",
       "clients_gender_0     2       11"
      ]
     },
     "execution_count": 201,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame({'test': (users_test[users_test.gender == 1].query(\"total_revenue != 0\").shape[0],\n",
    "                  users_test[users_test.gender == 0].query(\"total_revenue != 0\").shape[0]),\n",
    "              'control': (control_full[control_full.gender == 1].query(\"total_revenue != 0\").shape[0],\n",
    "                  control_full[control_full.gender == 0].query(\"total_revenue != 0\").shape[0])},\n",
    "              index=['clients_gender_1', 'clients_gender_0'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**У пользователей с гендером 0 средняя выручка упала, но общая выросла. Для дальнейшего анализа представляется недостаточное количество наблюдений (2 и 11 клиентов соответственно).**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LeveneResult(statistic=2.8122497712907233, pvalue=0.0378387352504359)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "levene(list(users_test[users_test.gender == 1].total_revenue),\n",
    "       list(users_test[users_test.gender == 0].total_revenue),\n",
    "       list(control_full[control_full.gender == 1].total_revenue),\n",
    "       list(control_full[control_full.gender == 0].total_revenue),\n",
    "       center='mean')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**С учетом того, что дисперсии внутри наших групп отличаются (требование гомогенности дисперсий не удовлетворено) целесообразно провести тест Крускала-Уоллиса.** "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KruskalResult(statistic=12.456189155495212, pvalue=0.005973163205387405)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.kruskal(users_test[users_test.gender == 1].total_revenue, users_test[users_test.gender == 0].total_revenue,\n",
    "    control_full[control_full.gender == 1].total_revenue, control_full[control_full.gender == 0].total_revenue)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Согласно значению pvalue нулевая гипотеза теста \"медиана одинакова для всех групп\" не может быть отвергнута.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_test = transactions_test.merge(users_test, on=['uid'], how='left')\n",
    "merged_control = control_transactions_full.merge(control_full, on=['uid'], how='left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KruskalResult(statistic=20.038378762655764, pvalue=0.0001666619546593342)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.kruskal(merged_test[merged_test.gender == 1].revenue,\n",
    "              merged_test[merged_test.gender == 0].revenue,\n",
    "              merged_control[merged_control.gender == 1].revenue,\n",
    "              merged_control[merged_control.gender == 0].revenue)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**В данных по транзакциям статистически значимые отличия имеются**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "p_values_gender = sp.posthoc_dunn([merged_test[merged_test.gender == 1].revenue,\n",
    "                                   merged_test[merged_test.gender == 0].revenue,\n",
    "                                   merged_control[merged_control.gender == 1].revenue,\n",
    "                                   merged_control[merged_control.gender == 0].revenue], p_adjust='bonferroni')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       1      2      3      4\n",
       "1  False  False   True  False\n",
       "2  False  False  False  False\n",
       "3   True  False  False  False\n",
       "4  False  False  False  False"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "p_values_gender < 0.05"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Отличия в выручке значимы только для гендера 1**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.2.2 Привлекательность<a name=\"attraction\"></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "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>Test</th>\n",
       "      <th>Control_1</th>\n",
       "      <th>Control_2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ex_premiums_on_uid</th>\n",
       "      <td>0.094708</td>\n",
       "      <td>0.100461</td>\n",
       "      <td>0.096388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>premiums_on_uid</th>\n",
       "      <td>0.036444</td>\n",
       "      <td>0.044240</td>\n",
       "      <td>0.044794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>revenue_on_uid</th>\n",
       "      <td>534.080316</td>\n",
       "      <td>594.761982</td>\n",
       "      <td>450.384146</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          Test   Control_1   Control_2\n",
       "ex_premiums_on_uid    0.094708    0.100461    0.096388\n",
       "premiums_on_uid       0.036444    0.044240    0.044794\n",
       "revenue_on_uid      534.080316  594.761982  450.384146"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "general_indicators"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SpearmanrResult(correlation=0.05513051498203405, pvalue=0.00029435922237452814) , SpearmanrResult(correlation=0.04848306859709989, pvalue=6.821842348228197e-06)\n"
     ]
    }
   ],
   "source": [
    "print(stats.spearmanr(list(users_test.attraction_coeff),\n",
    "                      list(users_test.total_revenue)), ',',\n",
    "      stats.spearmanr(list(control_full.attraction_coeff),\n",
    "                      list(control_full.total_revenue)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Распределения значений не являются нормальными, принимается решение о выборе коэффициента ранговой корреляции r-Спирмена. Выявлена значительная положительная корреляция между коэффициентом привлекательности и выручкой, которую можно не учитывать в силу величину коэффициента.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_merged_data = pd.concat([merged_test, merged_control], axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Выручка')"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(style='whitegrid', rc={'figure.figsize': (10, 5)})\n",
    "sns.scatterplot(x='attraction_coeff', y='revenue', data=full_merged_data)\n",
    "plt.title('Взаимосвязь между коэффициентом привлекательности и выручкой')\n",
    "plt.xlabel('Коэффициент привлекательности')\n",
    "plt.ylabel('Выручка')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SpearmanrResult(correlation=0.006456503565335202, pvalue=0.8401840513916059)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.spearmanr(list(full_merged_data.attraction_coeff),\n",
    "                list(full_merged_data.revenue))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.4     113.4\n",
       "0.7     500.0\n",
       "1.0    1000.0\n",
       "Name: attraction_coeff, dtype: float64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_data.attraction_coeff.quantile(q=[0.4, 0.7, 1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*В имеющихся данных большое количество пользователей с коэффициентом равным нулю,  методом подбора можно определить вышеуказанные границы групп пользователей*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "group1 = merged_test.query('attraction_coeff < 113').revenue \n",
    "group2 = merged_test.query('113 < attraction_coeff < 500').revenue \n",
    "group3 = merged_test.query('attraction_coeff > 500').revenue \n",
    "group4 = merged_control.query('attraction_coeff < 113').revenue \n",
    "group5 = merged_control.query('113 < attraction_coeff < 500').revenue \n",
    "group6 = merged_control.query('attraction_coeff > 500').revenue "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KruskalResult(statistic=26.09664700361864, pvalue=8.546583881232363e-05)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.kruskal(group1, group2, group3, group4, group5, group6)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Если результаты теста Краскела-Уоллиса статистически значимы, целесообразно провести тест Данна, чтобы точно определить, какие группы отличаются.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "p_values_coeff = sp.posthoc_dunn([group1, group2, group3, group4, group5, group6], p_adjust='bonferroni')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       1      2      3      4      5      6\n",
       "1  False  False  False  False  False  False\n",
       "2  False  False  False  False   True  False\n",
       "3  False  False  False  False  False  False\n",
       "4  False  False  False  False  False  False\n",
       "5  False   True  False  False  False   True\n",
       "6  False  False  False  False   True  False"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "p_values_coeff < 0.05"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**В представленной таблице нас интересуют пары 1-4, 2-5, 3-6. Для группы пользователей с коэффициентом привлекательности от 113 до 500 имеются статистически значимые различия.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.2.3 Количество посещений<a name=\"visits\"></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_for_split = full_data[['uid', 'visit_days', 'total_revenue', 'group']] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_for_split = data_for_split.fillna(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_for_split[\"total_days\"] = (data_for_split.visit_days.astype(str).apply(lambda x: x.split(',')[-1])).astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "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>uid</th>\n",
       "      <th>visit_days</th>\n",
       "      <th>total_revenue</th>\n",
       "      <th>group</th>\n",
       "      <th>total_days</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>892319115</td>\n",
       "      <td>1,2,3,4,5,6</td>\n",
       "      <td>0</td>\n",
       "      <td>control</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>891248523</td>\n",
       "      <td>1,2</td>\n",
       "      <td>0</td>\n",
       "      <td>control</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>891670932</td>\n",
       "      <td>1,2,3,5,6,8,9,11,13,16</td>\n",
       "      <td>0</td>\n",
       "      <td>control</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>891060786</td>\n",
       "      <td>1,2,3,4,5</td>\n",
       "      <td>0</td>\n",
       "      <td>control</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>892006554</td>\n",
       "      <td>1,2,3,5,6,7,8,9,10,12,15,16,17,19</td>\n",
       "      <td>0</td>\n",
       "      <td>control</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         uid                         visit_days  total_revenue    group  \\\n",
       "0  892319115                        1,2,3,4,5,6              0  control   \n",
       "1  891248523                                1,2              0  control   \n",
       "2  891670932             1,2,3,5,6,8,9,11,13,16              0  control   \n",
       "3  891060786                          1,2,3,4,5              0  control   \n",
       "4  892006554  1,2,3,5,6,7,8,9,10,12,15,16,17,19              0  control   \n",
       "\n",
       "   total_days  \n",
       "0           6  \n",
       "1           2  \n",
       "2          16  \n",
       "3           5  \n",
       "4          19  "
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_for_split.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SpearmanrResult(correlation=0.08801189806338636, pvalue=1.2531552460446705e-23)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.spearmanr(list(data_for_split.total_days),\n",
    "                list(data_for_split.total_revenue))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "avg_rev_per_days = data_for_split.groupby('total_days').agg({'total_revenue': 'mean'}) \\\n",
    "                                 .rename(columns={'total_revenue': 'avg_revenue'}) \\\n",
    "                                 .sort_values('avg_revenue', ascending=False) \\\n",
    "                                 .reset_index() \\\n",
    "                                 .head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "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>total_days</th>\n",
       "      <th>avg_revenue</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>28</td>\n",
       "      <td>4875.216667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15</td>\n",
       "      <td>3202.663043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>27</td>\n",
       "      <td>2146.054054</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>25</td>\n",
       "      <td>2009.913043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>60</td>\n",
       "      <td>1914.972222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>26</td>\n",
       "      <td>1817.531646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>29</td>\n",
       "      <td>1708.032258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>20</td>\n",
       "      <td>1611.878505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>24</td>\n",
       "      <td>1487.792079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>16</td>\n",
       "      <td>1337.562814</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  total_days  avg_revenue\n",
       "0         28  4875.216667\n",
       "1         15  3202.663043\n",
       "2         27  2146.054054\n",
       "3         25  2009.913043\n",
       "4         60  1914.972222\n",
       "5         26  1817.531646\n",
       "6         29  1708.032258\n",
       "7         20  1611.878505\n",
       "8         24  1487.792079\n",
       "9         16  1337.562814"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "avg_rev_per_days"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "27.0 , 12.6\n"
     ]
    }
   ],
   "source": [
    "print(avg_rev_per_days.total_days.astype('float64').mean().round(0), ',',\n",
    "      avg_rev_per_days.total_days.astype('float64').std().round(1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Для топ 10 групп пользователей по средней выручке, которые провели Х дней в приложении, среднее количество дней в приложении - 27. \\\n",
    "Стандартное отклонение при этом составило 12 дней. Таким образом стоит уделять внимание группе пользователей, посетивших приложение от 15 до 29 дней.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KruskalResult(statistic=149.80777008634732, pvalue=2.898903854513447e-32)"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.kruskal(data_for_split[data_for_split.group == 'control'].query('15 < total_days < 29').total_revenue,\n",
    "              data_for_split[data_for_split.group == 'test'].query('15 < total_days < 29').total_revenue,\n",
    "              data_for_split[data_for_split.group == 'control']\n",
    "                  .query('total_days < 15 | total_days > 29').total_revenue,\n",
    "              data_for_split[data_for_split.group == 'test']\n",
    "                  .query('total_days < 15 | total_days > 29').total_revenue)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "p_values_days = sp.posthoc_dunn([data_for_split[data_for_split.group == 'control']\n",
    "                                 .query('15 < total_days < 29').total_revenue,\n",
    "                                 data_for_split[data_for_split.group == 'test']\n",
    "                                 .query('15 < total_days < 29').total_revenue,\n",
    "                                 data_for_split[data_for_split.group == 'control']\n",
    "                                 .query('total_days < 15 | total_days > 29').total_revenue,\n",
    "                                 data_for_split[data_for_split.group == 'test']\n",
    "                                 .query('total_days < 15 | total_days > 29').total_revenue], p_adjust='bonferroni')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       1      2      3      4\n",
       "1  False   True   True   True\n",
       "2   True  False   True   True\n",
       "3   True   True  False  False\n",
       "4   True   True  False  False"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "p_values_days < 0.05"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Различия в выручке для выбранного диапазона посещаемости приложения значимы для тестовой и контрольных групп**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.2.4 Страны<a name=\"country\"></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_rev_avg = pd.concat([control_full.groupby('country').agg({'total_revenue': 'mean'})\n",
    "                                                   .sort_values('total_revenue', ascending=False),\n",
    "                                    users_test.groupby('country').agg({'total_revenue': 'mean'})\n",
    "                                                   .sort_values('total_revenue', ascending=False)],\n",
    "                                    axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_rev_avg.columns = ['control', 'test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_rev_avg['avg_rev_diff'] = country_rev_avg.test - country_rev_avg.control"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "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>control</th>\n",
       "      <th>test</th>\n",
       "      <th>rev_diff</th>\n",
       "      <th>avg_rev_diff</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Mexico</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>887.714286</td>\n",
       "      <td>887.714286</td>\n",
       "      <td>887.714286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Spain</th>\n",
       "      <td>449.138037</td>\n",
       "      <td>948.265781</td>\n",
       "      <td>499.127744</td>\n",
       "      <td>499.127744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>United Arab Emirates</th>\n",
       "      <td>460.971292</td>\n",
       "      <td>945.951327</td>\n",
       "      <td>484.980036</td>\n",
       "      <td>484.980036</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>France</th>\n",
       "      <td>446.949062</td>\n",
       "      <td>706.454039</td>\n",
       "      <td>259.504977</td>\n",
       "      <td>259.504977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Germany</th>\n",
       "      <td>244.633333</td>\n",
       "      <td>459.789474</td>\n",
       "      <td>215.156140</td>\n",
       "      <td>215.156140</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Argentina</th>\n",
       "      <td>338.049618</td>\n",
       "      <td>474.795455</td>\n",
       "      <td>136.745836</td>\n",
       "      <td>136.745836</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>United States of America</th>\n",
       "      <td>438.922557</td>\n",
       "      <td>497.819310</td>\n",
       "      <td>58.896754</td>\n",
       "      <td>58.896754</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Canada</th>\n",
       "      <td>457.425373</td>\n",
       "      <td>463.255474</td>\n",
       "      <td>5.830101</td>\n",
       "      <td>5.830101</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                             control        test    rev_diff  avg_rev_diff\n",
       "Mexico                      0.000000  887.714286  887.714286    887.714286\n",
       "Spain                     449.138037  948.265781  499.127744    499.127744\n",
       "United Arab Emirates      460.971292  945.951327  484.980036    484.980036\n",
       "France                    446.949062  706.454039  259.504977    259.504977\n",
       "Germany                   244.633333  459.789474  215.156140    215.156140\n",
       "Argentina                 338.049618  474.795455  136.745836    136.745836\n",
       "United States of America  438.922557  497.819310   58.896754     58.896754\n",
       "Canada                    457.425373  463.255474    5.830101      5.830101"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "country_rev_avg[country_rev_avg.rev_diff > 0].sort_values('avg_rev_diff', ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Необходимо проверить является ли статзначимой разница в распределении выручки за премиум подписки по странам Spain, UA Emirates, France, Germany, Argentina, USA.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_list = ['Spain', 'United Arab Emirates', 'France', 'Germany', 'Argentina', 'United States of America']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_df = full_merged_data.query(\"country_x == @country_list\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_group_1 = country_df[country_df.group == 'control'].query(\"country_x == 'Spain'\").revenue\n",
    "country_group_2 = country_df[country_df.group == 'control'].query(\"country_x == 'United Arab Emirates'\").revenue\n",
    "country_group_3 = country_df[country_df.group == 'control'].query(\"country_x == 'France'\").revenue\n",
    "country_group_4 = country_df[country_df.group == 'control'].query(\"country_x == 'Germany'\").revenue\n",
    "country_group_5 = country_df[country_df.group == 'control'].query(\"country_x == 'Argentina'\").revenue\n",
    "country_group_6 = country_df[country_df.group == 'control'].query(\"country_x == 'United States of America'\").revenue\n",
    "\n",
    "country_group_7 = country_df[country_df.group == 'test'].query(\"country_x == 'Spain'\").revenue\n",
    "country_group_8 = country_df[country_df.group == 'test'].query(\"country_x == 'United Arab Emirates'\").revenue\n",
    "country_group_9 = country_df[country_df.group == 'test'].query(\"country_x == 'France'\").revenue\n",
    "country_group_10 = country_df[country_df.group == 'test'].query(\"country_x == 'Germany'\").revenue\n",
    "country_group_11 = country_df[country_df.group == 'test'].query(\"country_x == 'Argentina'\").revenue\n",
    "country_group_12 = country_df[country_df.group == 'test'].query(\"country_x == 'United States of America'\").revenue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KruskalResult(statistic=34.12054072740245, pvalue=0.000345633734617267)"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.kruskal(country_group_1, country_group_2, country_group_3,\n",
    "              country_group_4, country_group_5, country_group_6,\n",
    "              country_group_7, country_group_8, country_group_9,\n",
    "              country_group_10, country_group_11, country_group_12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "country_values = sp.posthoc_dunn([country_group_1, country_group_2, country_group_3,\n",
    "                 country_group_4, country_group_5, country_group_6,\n",
    "                 country_group_7, country_group_8, country_group_9,\n",
    "                 country_group_10, country_group_11, country_group_12], p_adjust='bonferroni') < 0.05"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        1      2      3      4      5      6\n",
       "7   False  False  False  False  False  False\n",
       "8   False  False  False  False  False  False\n",
       "9   False  False  False  False  False  False\n",
       "10  False  False  False  False  False  False\n",
       "11  False  False   True  False   True   True\n",
       "12  False  False  False  False  False  False"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "country_values.iloc[6:12, 0:6]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**По результатам теста значимые отличия обнаружены только для страны Argentina**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.2.5 Возраст<a name=\"ages\"></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.052 , 0.31\n"
     ]
    }
   ],
   "source": [
    "print(stats.spearmanr(list(users_test.age),\n",
    "            list(users_test.total_revenue))[1].round(3), ',',\n",
    "      stats.spearmanr(list(control_full.age),\n",
    "            list(control_full.total_revenue))[1].round(3))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Коэффициенты корреляции между возрастом пользователя статистически не значимы**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Границы возрастного диапазона:  16 - 99\n"
     ]
    }
   ],
   "source": [
    "print('Границы возрастного диапазона: ', full_data.age.min(), '-', full_data.age.max())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "for_labels = ['16-20', '21-26', '27-32', '33-38', '39-46', '46-55', '56-99']\n",
    "for_bins = [15, 20, 26, 32, 38, 46, 55, 100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_merged_data = full_merged_data.assign(\n",
    "                   age_group=pd.cut(full_merged_data['age'], bins=for_bins, labels=for_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [],
   "source": [
    "formula = 'revenue ~ group + age_group + group: age_group'\n",
    "model = ols(formula, full_merged_data).fit()\n",
    "aov_table = anova_lm(model, typ=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                       sum_sq     df          F    PR(>F)\n",
      "group            1.024304e+09    1.0  13.829121  0.000212\n",
      "age_group        4.613758e+08    6.0   1.038172  0.398802\n",
      "group:age_group  4.177482e+08    6.0   0.940002  0.465265\n",
      "Residual         7.140219e+10  964.0        NaN       NaN\n"
     ]
    }
   ],
   "source": [
    "print(aov_table)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Как видно по последней колонке, эффект фактора age_group и взаимодействие факторов не значимы.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Выручка')"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(x='age_group', y='revenue', hue='group', data=full_merged_data)\n",
    "plt.xlabel('Возрастная группа')\n",
    "plt.ylabel('Выручка')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_group_1 = full_merged_data[full_merged_data.group == 'control'].query(\"age_group == '16-20'\").revenue\n",
    "age_group_2 = full_merged_data[full_merged_data.group == 'control'].query(\"age_group == '21-26'\").revenue\n",
    "age_group_3 = full_merged_data[full_merged_data.group == 'control'].query(\"age_group == '27-32'\").revenue\n",
    "age_group_4 = full_merged_data[full_merged_data.group == 'control'].query(\"age_group == '33-38'\").revenue\n",
    "age_group_5 = full_merged_data[full_merged_data.group == 'control'].query(\"age_group == '39-46'\").revenue\n",
    "age_group_6 = full_merged_data[full_merged_data.group == 'control'].query(\"age_group == '46-55'\").revenue\n",
    "age_group_7 = full_merged_data[full_merged_data.group == 'control'].query(\"age_group == '56-99'\").revenue\n",
    "\n",
    "age_group_8 = full_merged_data[full_merged_data.group == 'test'].query(\"age_group == '16-20'\").revenue\n",
    "age_group_9 = full_merged_data[full_merged_data.group == 'test'].query(\"age_group == '21-26'\").revenue\n",
    "age_group_10 = full_merged_data[full_merged_data.group == 'test'].query(\"age_group == '27-32'\").revenue\n",
    "age_group_11 = full_merged_data[full_merged_data.group == 'test'].query(\"age_group == '33-38'\").revenue\n",
    "age_group_12 = full_merged_data[full_merged_data.group == 'test'].query(\"age_group == '39-46'\").revenue\n",
    "age_group_13 = full_merged_data[full_merged_data.group == 'test'].query(\"age_group == '46-55'\").revenue\n",
    "age_group_14 = full_merged_data[full_merged_data.group == 'test'].query(\"age_group == '56-99'\").revenue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [],
   "source": [
    "age_pvalues = sp.posthoc_dunn([age_group_1, age_group_2, age_group_3, age_group_4, age_group_5, age_group_6,\n",
    "                               age_group_7, age_group_8, age_group_9, age_group_10, age_group_11, age_group_12,\n",
    "                               age_group_13, age_group_14], p_adjust='bonferroni') < 0.05"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        1      2      3      4      5      6      7\n",
       "8   False   True  False  False  False  False  False\n",
       "9   False  False  False  False  False  False  False\n",
       "10  False  False  False  False  False  False  False\n",
       "11  False  False  False  False  False  False  False\n",
       "12  False   True  False  False  False  False  False\n",
       "13  False  False  False  False  False  False  False\n",
       "14  False  False  False  False  False  False  False"
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "age_pvalues.iloc[7:14, 0:7]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**В выбранных возрастных группах статистически значимых отличий не обнаружено**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4. Выводы<a name=\"vivod\"></a>\n",
    "**В целом эксперимент был успешен. Для тестовой группы стоимость подписки была повышена (ARPPU увеличилась, количество премиум-подписок на пользователя сократилось), но изменение выручки от увеличения стоимости подписки больше, чем от сокращения количества клиентов.**\n",
    "\n",
    "**Нововведение имеет смысл для следующих групп:**\n",
    "   *  Пользователи гендера 1 \n",
    "   *  Посетившие приложение от 15 до 29 дней\n",
    "   *  Пользователи из страны Argentina\n",
    "   *  С коэффициентом привлекательности от 113 до 500"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}