11. User Referral Program.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "335d04e4",
   "metadata": {},
   "source": [
    "## Goal of Analysis:\n",
    "\n",
    "### 1. Evaluate the impact of the referral program\n",
    "\n",
    "### 2. What will be the next step based on the data\n",
    "\n",
    "### 3. Design a method to measure the result before and after the referral program"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75815332",
   "metadata": {},
   "source": [
    "### Part 1: Laod Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a81d2353",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "03af0589",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>date</th>\n",
       "      <th>country</th>\n",
       "      <th>money_spent</th>\n",
       "      <th>is_referral</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>FR</td>\n",
       "      <td>65</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>CA</td>\n",
       "      <td>54</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>FR</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>UK</td>\n",
       "      <td>73</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>MX</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id       date country  money_spent  is_referral\n",
       "0        2 2015-10-03      FR           65            0\n",
       "1        3 2015-10-03      CA           54            0\n",
       "2        6 2015-10-03      FR           35            0\n",
       "3        7 2015-10-03      UK           73            0\n",
       "4        7 2015-10-03      MX           35            0"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "referral = pd.read_csv('referral.csv', parse_dates = ['date'])\n",
    "\n",
    "# I will eliminate the device id for easier manipulation\n",
    "referral = referral[['user_id', 'date', 'country', 'money_spent', 'is_referral']]\n",
    "\n",
    "referral.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "169cb9b4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user_id                 int64\n",
       "date           datetime64[ns]\n",
       "country                object\n",
       "money_spent             int64\n",
       "is_referral             int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "referral.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d4a20f70",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user_id        0\n",
       "date           0\n",
       "country        0\n",
       "money_spent    0\n",
       "is_referral    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "referral.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2c060dee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'CA', 'CH', 'DE', 'ES', 'FR', 'IT', 'MX', 'UK', 'US'}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "set(referral['country'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a9888df2",
   "metadata": {},
   "source": [
    "I have converted the date to datetime and checked that the dataframe does not have null value"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "58e718d3",
   "metadata": {},
   "source": [
    "The program starts on Oct 31, 2015, so I need to check how many days before and after the program start date"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "110646e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import datetime\n",
    "\n",
    "referral_start_date = datetime.datetime(2015,10,31)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "1cc85948",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False    28\n",
       "True     28\n",
       "dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(pd.Series(referral.date.unique()) >= referral_start_date).value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5b9c9944",
   "metadata": {},
   "source": [
    "There are 28 days before and 28 days after the program start date, equally distributed"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d92a688",
   "metadata": {},
   "source": [
    "### Part 2: Data Preprocess and EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "6150db9f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>date</th>\n",
       "      <th>country</th>\n",
       "      <th>money_spent</th>\n",
       "      <th>is_referral</th>\n",
       "      <th>dayofweek</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>FR</td>\n",
       "      <td>65</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>CA</td>\n",
       "      <td>54</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>FR</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>UK</td>\n",
       "      <td>73</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7</td>\n",
       "      <td>2015-10-03</td>\n",
       "      <td>MX</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id       date country  money_spent  is_referral  dayofweek\n",
       "0        2 2015-10-03      FR           65            0          5\n",
       "1        3 2015-10-03      CA           54            0          5\n",
       "2        6 2015-10-03      FR           35            0          5\n",
       "3        7 2015-10-03      UK           73            0          5\n",
       "4        7 2015-10-03      MX           35            0          5"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# extract date information\n",
    "referral['dayofweek'] = referral['date'].apply(lambda x: x.dayofweek)\n",
    "referral.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4fc37e4d",
   "metadata": {},
   "source": [
    "We first take a look at the each country's referral and money spent information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "fdf32aa4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x864 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# daily money spent visualization\n",
    "fig, ax = plt.subplots(nrows=2, ncols=1, figsize=(18, 12))\n",
    "sns.countplot(x='country', hue='is_referral', data=referral, ax=ax[0])\n",
    "ax[0].set_ylabel('Country-wise Transaction Number', fontsize=12)\n",
    "\n",
    "sns.barplot(x='country', y='money_spent', hue='is_referral', data=referral, ax=ax[1])\n",
    "ax[1].set_ylabel('Country-wise Transaction Amount', fontsize=12)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd14d6c5",
   "metadata": {},
   "source": [
    "Also take a look at whether dayofweek has impact on the money spent on each country"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "59410db8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x864 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# daily money spent visualization (Monday is 0 and Sunday is 6)\n",
    "fig, ax = plt.subplots(nrows=2, ncols=1, figsize=(18, 12))\n",
    "sns.countplot(x='dayofweek', hue='is_referral', data=referral, ax=ax[0])\n",
    "ax[0].set_ylabel('Day-wise Transaction Number', fontsize=12)\n",
    "\n",
    "sns.barplot(x='dayofweek', y='money_spent', hue='is_referral', data=referral, ax=ax[1])\n",
    "ax[1].set_ylabel('Day-wise Transaction Amount', fontsize=12)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "050c859c",
   "metadata": {},
   "source": [
    "We can see that when there are referrals, people tend to spend more money. Also, people tend to spend more during weekends (day 5-6)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e8c881d7",
   "metadata": {},
   "source": [
    "### Part 3: Case Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "169a4aa0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def basic_stat(df):\n",
    "    \"\"\" function to calculate statistics \"\"\"\n",
    "    no_purchase = len(df)\n",
    "    money_spent = df['money_spent'].sum()\n",
    "    unique_user = len(df['user_id'].unique())\n",
    "    \n",
    "    return pd.Series([no_purchase, money_spent, unique_user], \n",
    "                     index=['no_purchase', 'money_spent', 'unique_user'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "bb746d96",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.stats import ttest_ind\n",
    "\n",
    "def daily_stat(data):\n",
    "    grouped = data.groupby('date').apply(basic_stat).reset_index()\n",
    "    ref = grouped[grouped['date'] < referral_start_date]\n",
    "    new = grouped[grouped['date'] >= referral_start_date]\n",
    "    \n",
    "    names = ['no_purchase', 'money_spent', 'unique_user']\n",
    "    vals = []\n",
    "    for name in names:\n",
    "        a = ref[name].values\n",
    "        b = new[name].values\n",
    "        # p-value should be one-side test\n",
    "        p_val = ttest_ind(a, b, equal_var=False)[1] / 2\n",
    "        vals.append(p_val)\n",
    "        \n",
    "    return pd.Series(vals, index=names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "a37a3f85",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "no_purchase    0.348257\n",
       "money_spent    0.135194\n",
       "unique_user    0.059545\n",
       "dtype: float64"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# analysis of user's daily information\n",
    "\n",
    "daily_stat(referral)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "0d62f057",
   "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>country</th>\n",
       "      <th>no_purchase</th>\n",
       "      <th>money_spent</th>\n",
       "      <th>unique_user</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CA</td>\n",
       "      <td>0.233985</td>\n",
       "      <td>0.351704</td>\n",
       "      <td>0.268256</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CH</td>\n",
       "      <td>0.003072</td>\n",
       "      <td>0.006941</td>\n",
       "      <td>0.003142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DE</td>\n",
       "      <td>0.011798</td>\n",
       "      <td>0.081459</td>\n",
       "      <td>0.015665</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ES</td>\n",
       "      <td>0.072638</td>\n",
       "      <td>0.037522</td>\n",
       "      <td>0.057954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>FR</td>\n",
       "      <td>0.058996</td>\n",
       "      <td>0.031843</td>\n",
       "      <td>0.041124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>IT</td>\n",
       "      <td>0.057454</td>\n",
       "      <td>0.025730</td>\n",
       "      <td>0.043911</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>MX</td>\n",
       "      <td>0.032430</td>\n",
       "      <td>0.009670</td>\n",
       "      <td>0.026203</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>UK</td>\n",
       "      <td>0.261183</td>\n",
       "      <td>0.048490</td>\n",
       "      <td>0.204398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>US</td>\n",
       "      <td>0.307801</td>\n",
       "      <td>0.248874</td>\n",
       "      <td>0.368860</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  country  no_purchase  money_spent  unique_user\n",
       "0      CA     0.233985     0.351704     0.268256\n",
       "1      CH     0.003072     0.006941     0.003142\n",
       "2      DE     0.011798     0.081459     0.015665\n",
       "3      ES     0.072638     0.037522     0.057954\n",
       "4      FR     0.058996     0.031843     0.041124\n",
       "5      IT     0.057454     0.025730     0.043911\n",
       "6      MX     0.032430     0.009670     0.026203\n",
       "7      UK     0.261183     0.048490     0.204398\n",
       "8      US     0.307801     0.248874     0.368860"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# analysis of country information\n",
    "\n",
    "referral.groupby('country').apply(daily_stat).reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "86129f64",
   "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>dayofweek</th>\n",
       "      <th>no_purchase</th>\n",
       "      <th>money_spent</th>\n",
       "      <th>unique_user</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.012484</td>\n",
       "      <td>0.000368</td>\n",
       "      <td>0.000665</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.156965</td>\n",
       "      <td>0.007701</td>\n",
       "      <td>0.008802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0.009713</td>\n",
       "      <td>0.000301</td>\n",
       "      <td>0.001270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.016584</td>\n",
       "      <td>0.000675</td>\n",
       "      <td>0.002578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0.016875</td>\n",
       "      <td>0.000595</td>\n",
       "      <td>0.000683</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>0.006828</td>\n",
       "      <td>0.000107</td>\n",
       "      <td>0.000125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>0.011717</td>\n",
       "      <td>0.000133</td>\n",
       "      <td>0.000012</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   dayofweek  no_purchase  money_spent  unique_user\n",
       "0          0     0.012484     0.000368     0.000665\n",
       "1          1     0.156965     0.007701     0.008802\n",
       "2          2     0.009713     0.000301     0.001270\n",
       "3          3     0.016584     0.000675     0.002578\n",
       "4          4     0.016875     0.000595     0.000683\n",
       "5          5     0.006828     0.000107     0.000125\n",
       "6          6     0.011717     0.000133     0.000012"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# analysis of dayofweek\n",
    "\n",
    "referral.groupby('dayofweek').apply(daily_stat).reset_index()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "630488c0",
   "metadata": {},
   "source": [
    "### Fianl Answer to the questions:\n",
    "\n",
    "#### 1. \n",
    "\n",
    "According to the analysis above, the program doesn't seem have significant impacts to the whole company as a whole.\n",
    "\n",
    "However, situations are different on each country:\n",
    "\n",
    " - the program fails in CH and DE, it significantly decrease the sales in these two countries.\n",
    " - the program succeeds in 'MX','IT','FR','ES', it significantly increase the sales.\n",
    " - the program doesn't seem have any significant effect on UK,CA,US, especially on CA and US.\n",
    "\n",
    "\n",
    "#### 2.\n",
    "\n",
    "I suggest to perform more accurate A/B test and collect more data, to study the impact of the program.\n",
    "\n",
    "since the program has different impact in different country, I suggest studying the reason of such difference. for example, does the program has any cultural conflicts in CH and DE?\n",
    "\n",
    "\n",
    "#### 3. \n",
    "\n",
    "This approach isn't an accurate A/B test. \"User Referral\" program isn't the only difference between control group and test group. For example, there may be some special holidays after Oct 31 in some countries that will influence the results.\n",
    "\n",
    "To get more accurate impact of the program, we need to perform a more careful A/B test. \n",
    "\n",
    "For example: during the same peroid of time, randomly split the customers into two groups, and let only one group know the User Referral program. Then run the experiment for a while and perform the t-test to see whether some 'sale performance index' (e.g., daily spent, daily customers, daily transactions) have significant changes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b8390926",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}