3_Clustering_kmeans.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clustering in Python using k-means"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Following notebook showcases the flow of a clustering exercise using customer sales data.\n",
    "\n",
    "It covers following steps:\n",
    "- Conversion of input sales data to a feature dataset that can be used for clustering\n",
    "- Performing clustering\n",
    "- Profiling the clusters, and\n",
    "- Setting up a regular scoring process to assign cluster labels basis new sales data.\n",
    "\n",
    "Python 3.7 has been used for this notebook. Environment package list is present in *requirements.txt*."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>The goal of this exercise is to group customers into *n* distinct clusters basis their sales activity and demographics data.</b>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Feature dataset creation\n",
    "We will explore the input transactional datasets and convert them to a customer level features dataset that can be used for the clustering exercise.\n",
    "\n",
    "<mark>Note</mark>: It's likely that the input transactional and other data that you want to use resides in a SQL database.<br> You can either connect your python code to the SQL database using relevant drivers and code and do the feature set creation in python, or preferably,<br> create the feature dataset in SQL and directly import it for clustering in python."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-03-30T12:51:35.259617Z",
     "start_time": "2021-03-30T12:51:35.246608Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\user\\anaconda3\\envs\\gen_ml\\lib\\site-packages\\statsmodels\\tools\\_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
      "  import pandas.util.testing as tm\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import pickle\n",
    "from tqdm import tqdm\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "from sklearn.metrics import silhouette_score as ss\n",
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-03-30T12:37:57.443440Z",
     "start_time": "2021-03-30T12:37:57.427447Z"
    }
   },
   "outputs": [],
   "source": [
    "pd.set_option(\"display.max_rows\", None, \"display.max_columns\", None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-03-30T12:37:58.182498Z",
     "start_time": "2021-03-30T12:37:58.166539Z"
    }
   },
   "outputs": [],
   "source": [
    "def show_null_count(df):\n",
    "    df_null = pd.DataFrame(data=df.isnull().sum(),columns=['nulls'])\n",
    "    df_null = df_null.sort_values('nulls', ascending=False)\n",
    "    return df_null"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Input data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Read all input datasets</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-03-30T12:39:01.032287Z",
     "start_time": "2021-03-30T12:38:19.459875Z"
    }
   },
   "outputs": [],
   "source": [
    "df_sales = pd.read_csv('./data/clustering_sales.csv')\n",
    "df_customer = pd.read_csv('./data/clustering_customer.csv')\n",
    "df_product = pd.read_csv('./data/clustering_product.csv')\n",
    "df_payment = pd.read_csv('./data/clustering_payment.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Quick exploration of the datasets:</b>\n",
    "1. We have sales data that captures date, customer id, product, quantity, dollar amount & payment type at order x item level. 'order_item_id' refers to each unique product in each order\n",
    "2. We have corresponding dimension tables for customer info, product info, and payment tender info"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>order_id</th>\n",
       "      <th>order_item_id</th>\n",
       "      <th>tran_dt</th>\n",
       "      <th>customer_id</th>\n",
       "      <th>dollars</th>\n",
       "      <th>qty</th>\n",
       "      <th>product_id</th>\n",
       "      <th>payment_type_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2020-01-01</td>\n",
       "      <td>572</td>\n",
       "      <td>550</td>\n",
       "      <td>1</td>\n",
       "      <td>20</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2020-01-01</td>\n",
       "      <td>532</td>\n",
       "      <td>630</td>\n",
       "      <td>3</td>\n",
       "      <td>11</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>2020-01-01</td>\n",
       "      <td>608</td>\n",
       "      <td>450</td>\n",
       "      <td>2</td>\n",
       "      <td>18</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>2020-01-01</td>\n",
       "      <td>424</td>\n",
       "      <td>110</td>\n",
       "      <td>2</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>2020-01-01</td>\n",
       "      <td>584</td>\n",
       "      <td>250</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   order_id  order_item_id     tran_dt  customer_id  dollars  qty  product_id  \\\n",
       "0         1              1  2020-01-01          572      550    1          20   \n",
       "1         2              2  2020-01-01          532      630    3          11   \n",
       "2         3              3  2020-01-01          608      450    2          18   \n",
       "3         4              4  2020-01-01          424      110    2          10   \n",
       "4         5              5  2020-01-01          584      250    1           8   \n",
       "\n",
       "   payment_type_id  \n",
       "0                2  \n",
       "1                2  \n",
       "2                4  \n",
       "3                2  \n",
       "4                4  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sales.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((10000, 8), 10000, 9811)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# order_item_id is the primary key\n",
    "(df_sales.shape, df_sales.order_item_id.nunique(), df_sales.order_id.nunique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 10000 entries, 0 to 9999\n",
      "Data columns (total 8 columns):\n",
      " #   Column           Non-Null Count  Dtype \n",
      "---  ------           --------------  ----- \n",
      " 0   order_id         10000 non-null  int64 \n",
      " 1   order_item_id    10000 non-null  int64 \n",
      " 2   tran_dt          10000 non-null  object\n",
      " 3   customer_id      10000 non-null  int64 \n",
      " 4   dollars          10000 non-null  int64 \n",
      " 5   qty              10000 non-null  int64 \n",
      " 6   product_id       10000 non-null  int64 \n",
      " 7   payment_type_id  10000 non-null  int64 \n",
      "dtypes: int64(7), object(1)\n",
      "memory usage: 625.1+ KB\n"
     ]
    }
   ],
   "source": [
    "df_sales.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customer_id</th>\n",
       "      <th>age</th>\n",
       "      <th>hh_income</th>\n",
       "      <th>omni_shopper</th>\n",
       "      <th>email_subscribed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>46</td>\n",
       "      <td>640000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>32</td>\n",
       "      <td>890000</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>45</td>\n",
       "      <td>772000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>46</td>\n",
       "      <td>303000</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>38</td>\n",
       "      <td>412000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id  age  hh_income  omni_shopper  email_subscribed\n",
       "0            1   46     640000             0                 0\n",
       "1            2   32     890000             1                 1\n",
       "2            3   45     772000             0                 0\n",
       "3            4   46     303000             0                 1\n",
       "4            5   38     412000             0                 0"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_customer.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1000, (1000, 5), 1000)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# we have 1k unique customers in sales data with all their info in customer dimension table\n",
    "(df_sales.customer_id.nunique(), df_customer.shape, df_customer.customer_id.nunique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>product_id</th>\n",
       "      <th>category</th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>A</td>\n",
       "      <td>450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>B</td>\n",
       "      <td>80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>C</td>\n",
       "      <td>250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>D</td>\n",
       "      <td>400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>E</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   product_id category  price\n",
       "0           1        A    450\n",
       "1           2        B     80\n",
       "2           3        C    250\n",
       "3           4        D    400\n",
       "4           5        E     50"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# product dimension table provides category and price for each product\n",
    "df_product.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((22, 3), 22)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(df_product.shape, df_product.product_id.nunique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>payment_type_id</th>\n",
       "      <th>payment_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>cash</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>credit card</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>debit card</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>gift card</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>others</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   payment_type_id payment_type\n",
       "0                1         cash\n",
       "1                2  credit card\n",
       "2                3   debit card\n",
       "3                4    gift card\n",
       "4                5       others"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# payment type table maps the payment type id from sales table\n",
    "df_payment.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Features dataset\n",
    "The goal of this exercise is to group customers into *n* distinct clusters basis their sales activity and demographics data.\n",
    "\n",
    "In order to cluster customers, we will require a dataset that summarizes their activity over time into a customer id level dataset, i.e. <b>we need to depict each customer using 1 row for data that covers everything we know about the customer</b>.\n",
    "\n",
    "The required output table should have 1 row per customer with relevant information as feature columns. There are many possible features that you can create here, some as simple as *sales $ for 1 year*, and some complicated or derived such as *quarterly change in sales of a customer for a particular product category*.\n",
    "\n",
    "This step of the process will involve conversations with the business stakeholders to determine what features matter to the business."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Features we will be creating for current exercise are listed below</b>\n",
    "\n",
    "Data we have is for the calendar year 2020 and we will take the whole time frame.<br>\n",
    "<mark>In your case, you might have to add relevant filters and join the sales dataset with some other datasets to get the required data that can then be aggregated to get below features.</mark>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Overall level:\n",
    "1. Sales\n",
    "2. Quantity\n",
    "3. No. of orders\n",
    "4. Avg. order value\n",
    "5. Units per transaction\n",
    "6. Avg. unit revenue\n",
    "7. No. of different products bought\n",
    "8. No. of different product categories bought\n",
    "9. No. of different payment types used"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Category level:\n",
    "1. Split of category level sales as % of total sales\n",
    "2. Split of category level units as % of total units"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Tender type level:\n",
    "1. Split of tender type level sales as % of total sales"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Customer information:\n",
    "1. Omni shopper flag\n",
    "2. Email subscription flag\n",
    "Other demographics data will be kept only for profiling."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Overall level features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sales = df_sales.merge(df_product[['product_id','category']], on=['product_id'])\n",
    "df_sales = df_sales.merge(df_payment, on=['payment_type_id'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features_overall = df_sales.groupby(['customer_id']).agg({\n",
    "    'dollars':'sum',\n",
    "    'qty':'sum',\n",
    "    'order_id':'nunique',\n",
    "    'product_id':'nunique',\n",
    "    'payment_type_id':'nunique',\n",
    "    'category':'nunique'\n",
    "    })"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features_overall['aov'] = df_features_overall['dollars']/df_features_overall['order_id']\n",
    "df_features_overall['aur'] = df_features_overall['dollars']/df_features_overall['qty']\n",
    "df_features_overall['upt'] = df_features_overall['qty']/df_features_overall['order_id']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features_overall.columns = [\n",
    "    'sales','units','orders','unique_products_bought','unique_payments_used',\n",
    "    'unique_categories_bought','aov','aur','upt']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sales</th>\n",
       "      <th>units</th>\n",
       "      <th>orders</th>\n",
       "      <th>unique_products_bought</th>\n",
       "      <th>unique_payments_used</th>\n",
       "      <th>unique_categories_bought</th>\n",
       "      <th>aov</th>\n",
       "      <th>aur</th>\n",
       "      <th>upt</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2395</td>\n",
       "      <td>11</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>479.000000</td>\n",
       "      <td>217.727273</td>\n",
       "      <td>2.200000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4815</td>\n",
       "      <td>15</td>\n",
       "      <td>7</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>687.857143</td>\n",
       "      <td>321.000000</td>\n",
       "      <td>2.142857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4285</td>\n",
       "      <td>21</td>\n",
       "      <td>10</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>428.500000</td>\n",
       "      <td>204.047619</td>\n",
       "      <td>2.100000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12000</td>\n",
       "      <td>44</td>\n",
       "      <td>21</td>\n",
       "      <td>15</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>571.428571</td>\n",
       "      <td>272.727273</td>\n",
       "      <td>2.095238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1700</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>566.666667</td>\n",
       "      <td>212.500000</td>\n",
       "      <td>2.666667</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             sales  units  orders  unique_products_bought  \\\n",
       "customer_id                                                 \n",
       "1             2395     11       5                       5   \n",
       "2             4815     15       7                       6   \n",
       "3             4285     21      10                       9   \n",
       "4            12000     44      21                      15   \n",
       "5             1700      8       3                       3   \n",
       "\n",
       "             unique_payments_used  unique_categories_bought         aov  \\\n",
       "customer_id                                                               \n",
       "1                               3                         3  479.000000   \n",
       "2                               4                         3  687.857143   \n",
       "3                               4                         5  428.500000   \n",
       "4                               4                         5  571.428571   \n",
       "5                               2                         2  566.666667   \n",
       "\n",
       "                    aur       upt  \n",
       "customer_id                        \n",
       "1            217.727273  2.200000  \n",
       "2            321.000000  2.142857  \n",
       "3            204.047619  2.100000  \n",
       "4            272.727273  2.095238  \n",
       "5            212.500000  2.666667  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features_overall.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Category level features\n",
    "<b>Sales</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_category_features_s = df_sales.groupby(['customer_id','category']).agg({'dollars':'sum'}).reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# add overall sales\n",
    "df_category_features_s = df_category_features_s.merge(df_features_overall[['sales']], on=['customer_id'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert to %\n",
    "df_category_features_s['sales_perc'] = df_category_features_s['dollars']/df_category_features_s['sales']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customer_id</th>\n",
       "      <th>category</th>\n",
       "      <th>dollars</th>\n",
       "      <th>sales</th>\n",
       "      <th>sales_perc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>A</td>\n",
       "      <td>1150</td>\n",
       "      <td>2395</td>\n",
       "      <td>0.480167</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>C</td>\n",
       "      <td>1080</td>\n",
       "      <td>2395</td>\n",
       "      <td>0.450939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>E</td>\n",
       "      <td>165</td>\n",
       "      <td>2395</td>\n",
       "      <td>0.068894</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>A</td>\n",
       "      <td>3475</td>\n",
       "      <td>4815</td>\n",
       "      <td>0.721703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>C</td>\n",
       "      <td>1190</td>\n",
       "      <td>4815</td>\n",
       "      <td>0.247144</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id category  dollars  sales  sales_perc\n",
       "0            1        A     1150   2395    0.480167\n",
       "1            1        C     1080   2395    0.450939\n",
       "2            1        E      165   2395    0.068894\n",
       "3            2        A     3475   4815    0.721703\n",
       "4            2        C     1190   4815    0.247144"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_category_features_s.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# pivot\n",
    "df_category_features_s = df_category_features_s.pivot(index='customer_id', columns='category', values='sales_perc')\n",
    "df_category_features_s.columns = [\n",
    "    'category_a_sales','category_b_sales','category_c_sales','category_d_sales','category_e_sales']"
   ]
  },
  {
   "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>category_a_sales</th>\n",
       "      <th>category_b_sales</th>\n",
       "      <th>category_c_sales</th>\n",
       "      <th>category_d_sales</th>\n",
       "      <th>category_e_sales</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.480167</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.450939</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.068894</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.721703</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.247144</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.031153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.240373</td>\n",
       "      <td>0.175029</td>\n",
       "      <td>0.309218</td>\n",
       "      <td>0.210035</td>\n",
       "      <td>0.065344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.575000</td>\n",
       "      <td>0.067500</td>\n",
       "      <td>0.158333</td>\n",
       "      <td>0.170833</td>\n",
       "      <td>0.028333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.735294</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             category_a_sales  category_b_sales  category_c_sales  \\\n",
       "customer_id                                                         \n",
       "1                    0.480167               NaN          0.450939   \n",
       "2                    0.721703               NaN          0.247144   \n",
       "3                    0.240373          0.175029          0.309218   \n",
       "4                    0.575000          0.067500          0.158333   \n",
       "5                         NaN          0.264706          0.735294   \n",
       "\n",
       "             category_d_sales  category_e_sales  \n",
       "customer_id                                      \n",
       "1                         NaN          0.068894  \n",
       "2                         NaN          0.031153  \n",
       "3                    0.210035          0.065344  \n",
       "4                    0.170833          0.028333  \n",
       "5                         NaN               NaN  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_category_features_s.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Units</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_category_features_u = df_sales.groupby(['customer_id','category']).agg({'qty':'sum'}).reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# add overall sales\n",
    "df_category_features_u = df_category_features_u.merge(df_features_overall[['units']], on=['customer_id'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert to %\n",
    "df_category_features_u['units_perc'] = df_category_features_u['qty']/df_category_features_u['units']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customer_id</th>\n",
       "      <th>category</th>\n",
       "      <th>qty</th>\n",
       "      <th>units</th>\n",
       "      <th>units_perc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>A</td>\n",
       "      <td>4</td>\n",
       "      <td>11</td>\n",
       "      <td>0.363636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>C</td>\n",
       "      <td>4</td>\n",
       "      <td>11</td>\n",
       "      <td>0.363636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>E</td>\n",
       "      <td>3</td>\n",
       "      <td>11</td>\n",
       "      <td>0.272727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>A</td>\n",
       "      <td>7</td>\n",
       "      <td>15</td>\n",
       "      <td>0.466667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>C</td>\n",
       "      <td>5</td>\n",
       "      <td>15</td>\n",
       "      <td>0.333333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id category  qty  units  units_perc\n",
       "0            1        A    4     11    0.363636\n",
       "1            1        C    4     11    0.363636\n",
       "2            1        E    3     11    0.272727\n",
       "3            2        A    7     15    0.466667\n",
       "4            2        C    5     15    0.333333"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_category_features_u.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# pivot\n",
    "df_category_features_u = df_category_features_u.pivot(index='customer_id', columns='category', values='units_perc')\n",
    "df_category_features_u.columns = [\n",
    "    'category_a_units','category_b_units','category_c_units','category_d_units','category_e_units']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>category_a_units</th>\n",
       "      <th>category_b_units</th>\n",
       "      <th>category_c_units</th>\n",
       "      <th>category_d_units</th>\n",
       "      <th>category_e_units</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.363636</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.272727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.466667</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.200000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.190476</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.095238</td>\n",
       "      <td>0.238095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.409091</td>\n",
       "      <td>0.159091</td>\n",
       "      <td>0.181818</td>\n",
       "      <td>0.113636</td>\n",
       "      <td>0.136364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             category_a_units  category_b_units  category_c_units  \\\n",
       "customer_id                                                         \n",
       "1                    0.363636               NaN          0.363636   \n",
       "2                    0.466667               NaN          0.333333   \n",
       "3                    0.190476          0.238095          0.238095   \n",
       "4                    0.409091          0.159091          0.181818   \n",
       "5                         NaN          0.375000          0.625000   \n",
       "\n",
       "             category_d_units  category_e_units  \n",
       "customer_id                                      \n",
       "1                         NaN          0.272727  \n",
       "2                         NaN          0.200000  \n",
       "3                    0.095238          0.238095  \n",
       "4                    0.113636          0.136364  \n",
       "5                         NaN               NaN  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_category_features_u.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Tender type level"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_payment_features = df_sales.groupby(['customer_id','payment_type']).agg({'dollars':'sum'}).reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# add overall sales\n",
    "df_payment_features = df_payment_features.merge(df_features_overall[['sales']], on=['customer_id'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert to %\n",
    "df_payment_features['sales_perc'] = df_payment_features['dollars']/df_payment_features['sales']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>customer_id</th>\n",
       "      <th>payment_type</th>\n",
       "      <th>dollars</th>\n",
       "      <th>sales</th>\n",
       "      <th>sales_perc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>credit card</td>\n",
       "      <td>1245</td>\n",
       "      <td>2395</td>\n",
       "      <td>0.519833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>debit card</td>\n",
       "      <td>400</td>\n",
       "      <td>2395</td>\n",
       "      <td>0.167015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>gift card</td>\n",
       "      <td>750</td>\n",
       "      <td>2395</td>\n",
       "      <td>0.313152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>cash</td>\n",
       "      <td>150</td>\n",
       "      <td>4815</td>\n",
       "      <td>0.031153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>credit card</td>\n",
       "      <td>3390</td>\n",
       "      <td>4815</td>\n",
       "      <td>0.704050</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   customer_id payment_type  dollars  sales  sales_perc\n",
       "0            1  credit card     1245   2395    0.519833\n",
       "1            1   debit card      400   2395    0.167015\n",
       "2            1    gift card      750   2395    0.313152\n",
       "3            2         cash      150   4815    0.031153\n",
       "4            2  credit card     3390   4815    0.704050"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_payment_features.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# pivot\n",
    "df_payment_features = df_payment_features.pivot(index='customer_id', columns='payment_type', values='sales_perc')\n",
    "df_payment_features.columns = [\n",
    "    'payment_cash','payment_credit','payment_debit','payment_gc','payment_others']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>payment_cash</th>\n",
       "      <th>payment_credit</th>\n",
       "      <th>payment_debit</th>\n",
       "      <th>payment_gc</th>\n",
       "      <th>payment_others</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.519833</td>\n",
       "      <td>0.167015</td>\n",
       "      <td>0.313152</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.031153</td>\n",
       "      <td>0.704050</td>\n",
       "      <td>0.186916</td>\n",
       "      <td>0.077882</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.263711</td>\n",
       "      <td>0.309218</td>\n",
       "      <td>0.147025</td>\n",
       "      <td>0.280047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.345000</td>\n",
       "      <td>0.370000</td>\n",
       "      <td>0.218333</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.735294</td>\n",
       "      <td>0.264706</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             payment_cash  payment_credit  payment_debit  payment_gc  \\\n",
       "customer_id                                                            \n",
       "1                     NaN        0.519833       0.167015    0.313152   \n",
       "2                0.031153        0.704050       0.186916    0.077882   \n",
       "3                     NaN        0.263711       0.309218    0.147025   \n",
       "4                0.066667        0.345000       0.370000    0.218333   \n",
       "5                     NaN             NaN       0.735294    0.264706   \n",
       "\n",
       "             payment_others  \n",
       "customer_id                  \n",
       "1                       NaN  \n",
       "2                       NaN  \n",
       "3                  0.280047  \n",
       "4                       NaN  \n",
       "5                       NaN  "
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_payment_features.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Final Features dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features = df_features_overall.merge(df_category_features_s,on='customer_id',how='left')\n",
    "\n",
    "df_features = df_features.merge(df_category_features_u,on='customer_id',how='left')\n",
    "\n",
    "df_features = df_features.merge(df_payment_features,on='customer_id',how='left')\n",
    "\n",
    "df_features = df_features.merge(\n",
    "    df_customer[['customer_id','email_subscribed','omni_shopper']].set_index('customer_id'),\n",
    "    on='customer_id',how='left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>sales</th>\n",
       "      <th>units</th>\n",
       "      <th>orders</th>\n",
       "      <th>unique_products_bought</th>\n",
       "      <th>unique_payments_used</th>\n",
       "      <th>unique_categories_bought</th>\n",
       "      <th>aov</th>\n",
       "      <th>aur</th>\n",
       "      <th>upt</th>\n",
       "      <th>category_a_sales</th>\n",
       "      <th>category_b_sales</th>\n",
       "      <th>category_c_sales</th>\n",
       "      <th>category_d_sales</th>\n",
       "      <th>category_e_sales</th>\n",
       "      <th>category_a_units</th>\n",
       "      <th>category_b_units</th>\n",
       "      <th>category_c_units</th>\n",
       "      <th>category_d_units</th>\n",
       "      <th>category_e_units</th>\n",
       "      <th>payment_cash</th>\n",
       "      <th>payment_credit</th>\n",
       "      <th>payment_debit</th>\n",
       "      <th>payment_gc</th>\n",
       "      <th>payment_others</th>\n",
       "      <th>email_subscribed</th>\n",
       "      <th>omni_shopper</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2395</td>\n",
       "      <td>11</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>479.000000</td>\n",
       "      <td>217.727273</td>\n",
       "      <td>2.200000</td>\n",
       "      <td>0.480167</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.450939</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.068894</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.272727</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.519833</td>\n",
       "      <td>0.167015</td>\n",
       "      <td>0.313152</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4815</td>\n",
       "      <td>15</td>\n",
       "      <td>7</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>687.857143</td>\n",
       "      <td>321.000000</td>\n",
       "      <td>2.142857</td>\n",
       "      <td>0.721703</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.247144</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.031153</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.031153</td>\n",
       "      <td>0.704050</td>\n",
       "      <td>0.186916</td>\n",
       "      <td>0.077882</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4285</td>\n",
       "      <td>21</td>\n",
       "      <td>10</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>428.500000</td>\n",
       "      <td>204.047619</td>\n",
       "      <td>2.100000</td>\n",
       "      <td>0.240373</td>\n",
       "      <td>0.175029</td>\n",
       "      <td>0.309218</td>\n",
       "      <td>0.210035</td>\n",
       "      <td>0.065344</td>\n",
       "      <td>0.190476</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.095238</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.263711</td>\n",
       "      <td>0.309218</td>\n",
       "      <td>0.147025</td>\n",
       "      <td>0.280047</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12000</td>\n",
       "      <td>44</td>\n",
       "      <td>21</td>\n",
       "      <td>15</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>571.428571</td>\n",
       "      <td>272.727273</td>\n",
       "      <td>2.095238</td>\n",
       "      <td>0.575000</td>\n",
       "      <td>0.067500</td>\n",
       "      <td>0.158333</td>\n",
       "      <td>0.170833</td>\n",
       "      <td>0.028333</td>\n",
       "      <td>0.409091</td>\n",
       "      <td>0.159091</td>\n",
       "      <td>0.181818</td>\n",
       "      <td>0.113636</td>\n",
       "      <td>0.136364</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.345000</td>\n",
       "      <td>0.370000</td>\n",
       "      <td>0.218333</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1700</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>566.666667</td>\n",
       "      <td>212.500000</td>\n",
       "      <td>2.666667</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.735294</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.735294</td>\n",
       "      <td>0.264706</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             sales  units  orders  unique_products_bought  \\\n",
       "customer_id                                                 \n",
       "1             2395     11       5                       5   \n",
       "2             4815     15       7                       6   \n",
       "3             4285     21      10                       9   \n",
       "4            12000     44      21                      15   \n",
       "5             1700      8       3                       3   \n",
       "\n",
       "             unique_payments_used  unique_categories_bought         aov  \\\n",
       "customer_id                                                               \n",
       "1                               3                         3  479.000000   \n",
       "2                               4                         3  687.857143   \n",
       "3                               4                         5  428.500000   \n",
       "4                               4                         5  571.428571   \n",
       "5                               2                         2  566.666667   \n",
       "\n",
       "                    aur       upt  category_a_sales  category_b_sales  \\\n",
       "customer_id                                                             \n",
       "1            217.727273  2.200000          0.480167               NaN   \n",
       "2            321.000000  2.142857          0.721703               NaN   \n",
       "3            204.047619  2.100000          0.240373          0.175029   \n",
       "4            272.727273  2.095238          0.575000          0.067500   \n",
       "5            212.500000  2.666667               NaN          0.264706   \n",
       "\n",
       "             category_c_sales  category_d_sales  category_e_sales  \\\n",
       "customer_id                                                         \n",
       "1                    0.450939               NaN          0.068894   \n",
       "2                    0.247144               NaN          0.031153   \n",
       "3                    0.309218          0.210035          0.065344   \n",
       "4                    0.158333          0.170833          0.028333   \n",
       "5                    0.735294               NaN               NaN   \n",
       "\n",
       "             category_a_units  category_b_units  category_c_units  \\\n",
       "customer_id                                                         \n",
       "1                    0.363636               NaN          0.363636   \n",
       "2                    0.466667               NaN          0.333333   \n",
       "3                    0.190476          0.238095          0.238095   \n",
       "4                    0.409091          0.159091          0.181818   \n",
       "5                         NaN          0.375000          0.625000   \n",
       "\n",
       "             category_d_units  category_e_units  payment_cash  payment_credit  \\\n",
       "customer_id                                                                     \n",
       "1                         NaN          0.272727           NaN        0.519833   \n",
       "2                         NaN          0.200000      0.031153        0.704050   \n",
       "3                    0.095238          0.238095           NaN        0.263711   \n",
       "4                    0.113636          0.136364      0.066667        0.345000   \n",
       "5                         NaN               NaN           NaN             NaN   \n",
       "\n",
       "             payment_debit  payment_gc  payment_others  email_subscribed  \\\n",
       "customer_id                                                                \n",
       "1                 0.167015    0.313152             NaN                 0   \n",
       "2                 0.186916    0.077882             NaN                 1   \n",
       "3                 0.309218    0.147025        0.280047                 0   \n",
       "4                 0.370000    0.218333             NaN                 1   \n",
       "5                 0.735294    0.264706             NaN                 0   \n",
       "\n",
       "             omni_shopper  \n",
       "customer_id                \n",
       "1                       0  \n",
       "2                       1  \n",
       "3                       0  \n",
       "4                       0  \n",
       "5                       0  "
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features.to_csv('./data/clustering_features.csv', index=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. k-means clustering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary of features\n",
    "\n",
    "- There are nulls in category and payment % split features: fill with 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>sales</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>5707.170000</td>\n",
       "      <td>4036.079859</td>\n",
       "      <td>330.000000</td>\n",
       "      <td>2827.500000</td>\n",
       "      <td>4280.000000</td>\n",
       "      <td>7267.500000</td>\n",
       "      <td>17635.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>units</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>21.108000</td>\n",
       "      <td>14.408398</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>23.000000</td>\n",
       "      <td>58.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>orders</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>9.811000</td>\n",
       "      <td>6.384986</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>24.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_products_bought</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>7.547000</td>\n",
       "      <td>3.850535</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>18.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_payments_used</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>3.005000</td>\n",
       "      <td>0.948617</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>5.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_categories_bought</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>3.843000</td>\n",
       "      <td>1.014584</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aov</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>575.336384</td>\n",
       "      <td>165.741267</td>\n",
       "      <td>110.000000</td>\n",
       "      <td>467.500000</td>\n",
       "      <td>568.444444</td>\n",
       "      <td>670.238095</td>\n",
       "      <td>1202.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aur</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>271.150530</td>\n",
       "      <td>65.941495</td>\n",
       "      <td>60.000000</td>\n",
       "      <td>227.836538</td>\n",
       "      <td>268.178138</td>\n",
       "      <td>311.271777</td>\n",
       "      <td>485.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>upt</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>2.124689</td>\n",
       "      <td>0.344007</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.900000</td>\n",
       "      <td>2.139610</td>\n",
       "      <td>2.333333</td>\n",
       "      <td>3.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_a_sales</th>\n",
       "      <td>911.0</td>\n",
       "      <td>0.465008</td>\n",
       "      <td>0.200729</td>\n",
       "      <td>0.040698</td>\n",
       "      <td>0.316973</td>\n",
       "      <td>0.459397</td>\n",
       "      <td>0.598472</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_b_sales</th>\n",
       "      <td>754.0</td>\n",
       "      <td>0.129013</td>\n",
       "      <td>0.109362</td>\n",
       "      <td>0.005324</td>\n",
       "      <td>0.059289</td>\n",
       "      <td>0.103094</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_c_sales</th>\n",
       "      <td>829.0</td>\n",
       "      <td>0.275343</td>\n",
       "      <td>0.169208</td>\n",
       "      <td>0.021024</td>\n",
       "      <td>0.154321</td>\n",
       "      <td>0.244739</td>\n",
       "      <td>0.352052</td>\n",
       "      <td>0.951860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_d_sales</th>\n",
       "      <td>678.0</td>\n",
       "      <td>0.317048</td>\n",
       "      <td>0.173845</td>\n",
       "      <td>0.033003</td>\n",
       "      <td>0.183344</td>\n",
       "      <td>0.281823</td>\n",
       "      <td>0.422400</td>\n",
       "      <td>0.940171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_e_sales</th>\n",
       "      <td>671.0</td>\n",
       "      <td>0.053478</td>\n",
       "      <td>0.064436</td>\n",
       "      <td>0.003299</td>\n",
       "      <td>0.022009</td>\n",
       "      <td>0.037415</td>\n",
       "      <td>0.062500</td>\n",
       "      <td>0.809524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_a_units</th>\n",
       "      <td>911.0</td>\n",
       "      <td>0.351488</td>\n",
       "      <td>0.173091</td>\n",
       "      <td>0.043478</td>\n",
       "      <td>0.226415</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_b_units</th>\n",
       "      <td>754.0</td>\n",
       "      <td>0.241836</td>\n",
       "      <td>0.134064</td>\n",
       "      <td>0.023256</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>0.218452</td>\n",
       "      <td>0.312500</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_c_units</th>\n",
       "      <td>829.0</td>\n",
       "      <td>0.271817</td>\n",
       "      <td>0.149078</td>\n",
       "      <td>0.022727</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.857143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_d_units</th>\n",
       "      <td>678.0</td>\n",
       "      <td>0.200435</td>\n",
       "      <td>0.126031</td>\n",
       "      <td>0.019608</td>\n",
       "      <td>0.105263</td>\n",
       "      <td>0.168699</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.727273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_e_units</th>\n",
       "      <td>671.0</td>\n",
       "      <td>0.203009</td>\n",
       "      <td>0.131912</td>\n",
       "      <td>0.019608</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.263158</td>\n",
       "      <td>0.888889</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_cash</th>\n",
       "      <td>263.0</td>\n",
       "      <td>0.128239</td>\n",
       "      <td>0.126068</td>\n",
       "      <td>0.003754</td>\n",
       "      <td>0.038977</td>\n",
       "      <td>0.095969</td>\n",
       "      <td>0.168276</td>\n",
       "      <td>0.947867</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_credit</th>\n",
       "      <td>981.0</td>\n",
       "      <td>0.552566</td>\n",
       "      <td>0.211687</td>\n",
       "      <td>0.031339</td>\n",
       "      <td>0.411765</td>\n",
       "      <td>0.544426</td>\n",
       "      <td>0.702323</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_debit</th>\n",
       "      <td>871.0</td>\n",
       "      <td>0.307976</td>\n",
       "      <td>0.183177</td>\n",
       "      <td>0.008929</td>\n",
       "      <td>0.172142</td>\n",
       "      <td>0.277212</td>\n",
       "      <td>0.406788</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_gc</th>\n",
       "      <td>634.0</td>\n",
       "      <td>0.202083</td>\n",
       "      <td>0.160023</td>\n",
       "      <td>0.007018</td>\n",
       "      <td>0.081417</td>\n",
       "      <td>0.157143</td>\n",
       "      <td>0.275592</td>\n",
       "      <td>0.968661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_others</th>\n",
       "      <td>256.0</td>\n",
       "      <td>0.108743</td>\n",
       "      <td>0.102174</td>\n",
       "      <td>0.003154</td>\n",
       "      <td>0.041317</td>\n",
       "      <td>0.080988</td>\n",
       "      <td>0.144144</td>\n",
       "      <td>0.605042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>email_subscribed</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.624000</td>\n",
       "      <td>0.484622</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>omni_shopper</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.249000</td>\n",
       "      <td>0.432650</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           count         mean          std         min  \\\n",
       "sales                     1000.0  5707.170000  4036.079859  330.000000   \n",
       "units                     1000.0    21.108000    14.408398    3.000000   \n",
       "orders                    1000.0     9.811000     6.384986    3.000000   \n",
       "unique_products_bought    1000.0     7.547000     3.850535    2.000000   \n",
       "unique_payments_used      1000.0     3.005000     0.948617    1.000000   \n",
       "unique_categories_bought  1000.0     3.843000     1.014584    1.000000   \n",
       "aov                       1000.0   575.336384   165.741267  110.000000   \n",
       "aur                       1000.0   271.150530    65.941495   60.000000   \n",
       "upt                       1000.0     2.124689     0.344007    1.000000   \n",
       "category_a_sales           911.0     0.465008     0.200729    0.040698   \n",
       "category_b_sales           754.0     0.129013     0.109362    0.005324   \n",
       "category_c_sales           829.0     0.275343     0.169208    0.021024   \n",
       "category_d_sales           678.0     0.317048     0.173845    0.033003   \n",
       "category_e_sales           671.0     0.053478     0.064436    0.003299   \n",
       "category_a_units           911.0     0.351488     0.173091    0.043478   \n",
       "category_b_units           754.0     0.241836     0.134064    0.023256   \n",
       "category_c_units           829.0     0.271817     0.149078    0.022727   \n",
       "category_d_units           678.0     0.200435     0.126031    0.019608   \n",
       "category_e_units           671.0     0.203009     0.131912    0.019608   \n",
       "payment_cash               263.0     0.128239     0.126068    0.003754   \n",
       "payment_credit             981.0     0.552566     0.211687    0.031339   \n",
       "payment_debit              871.0     0.307976     0.183177    0.008929   \n",
       "payment_gc                 634.0     0.202083     0.160023    0.007018   \n",
       "payment_others             256.0     0.108743     0.102174    0.003154   \n",
       "email_subscribed          1000.0     0.624000     0.484622    0.000000   \n",
       "omni_shopper              1000.0     0.249000     0.432650    0.000000   \n",
       "\n",
       "                                  25%          50%          75%           max  \n",
       "sales                     2827.500000  4280.000000  7267.500000  17635.000000  \n",
       "units                       11.000000    16.000000    23.000000     58.000000  \n",
       "orders                       5.000000     8.000000    10.000000     24.000000  \n",
       "unique_products_bought       5.000000     6.000000     9.000000     18.000000  \n",
       "unique_payments_used         2.000000     3.000000     4.000000      5.000000  \n",
       "unique_categories_bought     3.000000     4.000000     5.000000      5.000000  \n",
       "aov                        467.500000   568.444444   670.238095   1202.500000  \n",
       "aur                        227.836538   268.178138   311.271777    485.000000  \n",
       "upt                          1.900000     2.139610     2.333333      3.666667  \n",
       "category_a_sales             0.316973     0.459397     0.598472      1.000000  \n",
       "category_b_sales             0.059289     0.103094     0.166667      1.000000  \n",
       "category_c_sales             0.154321     0.244739     0.352052      0.951860  \n",
       "category_d_sales             0.183344     0.281823     0.422400      0.940171  \n",
       "category_e_sales             0.022009     0.037415     0.062500      0.809524  \n",
       "category_a_units             0.226415     0.333333     0.454545      1.000000  \n",
       "category_b_units             0.142857     0.218452     0.312500      1.000000  \n",
       "category_c_units             0.166667     0.250000     0.333333      0.857143  \n",
       "category_d_units             0.105263     0.168699     0.266667      0.727273  \n",
       "category_e_units             0.111111     0.166667     0.263158      0.888889  \n",
       "payment_cash                 0.038977     0.095969     0.168276      0.947867  \n",
       "payment_credit               0.411765     0.544426     0.702323      1.000000  \n",
       "payment_debit                0.172142     0.277212     0.406788      1.000000  \n",
       "payment_gc                   0.081417     0.157143     0.275592      0.968661  \n",
       "payment_others               0.041317     0.080988     0.144144      0.605042  \n",
       "email_subscribed             0.000000     1.000000     1.000000      1.000000  \n",
       "omni_shopper                 0.000000     0.000000     0.000000      1.000000  "
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features.describe().T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "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>nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>payment_others</th>\n",
       "      <td>744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_cash</th>\n",
       "      <td>737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_gc</th>\n",
       "      <td>366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_e_sales</th>\n",
       "      <td>329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_e_units</th>\n",
       "      <td>329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_d_sales</th>\n",
       "      <td>322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_d_units</th>\n",
       "      <td>322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_b_sales</th>\n",
       "      <td>246</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_b_units</th>\n",
       "      <td>246</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_c_units</th>\n",
       "      <td>171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_c_sales</th>\n",
       "      <td>171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_debit</th>\n",
       "      <td>129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_a_sales</th>\n",
       "      <td>89</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_a_units</th>\n",
       "      <td>89</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_credit</th>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>email_subscribed</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sales</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>units</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>upt</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aur</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aov</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_categories_bought</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_payments_used</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_products_bought</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>orders</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>omni_shopper</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          nulls\n",
       "payment_others              744\n",
       "payment_cash                737\n",
       "payment_gc                  366\n",
       "category_e_sales            329\n",
       "category_e_units            329\n",
       "category_d_sales            322\n",
       "category_d_units            322\n",
       "category_b_sales            246\n",
       "category_b_units            246\n",
       "category_c_units            171\n",
       "category_c_sales            171\n",
       "payment_debit               129\n",
       "category_a_sales             89\n",
       "category_a_units             89\n",
       "payment_credit               19\n",
       "email_subscribed              0\n",
       "sales                         0\n",
       "units                         0\n",
       "upt                           0\n",
       "aur                           0\n",
       "aov                           0\n",
       "unique_categories_bought      0\n",
       "unique_payments_used          0\n",
       "unique_products_bought        0\n",
       "orders                        0\n",
       "omni_shopper                  0"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show_null_count(df_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features = df_features.fillna(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_features.isnull().sum().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature scaling\n",
    "\n",
    "- We will use MinMaxScaler for all non-% features\n",
    "- % columns are already in \\[0,1\\] and hence scaling is not required\n",
    "- *email_subscribed* and *omni_shopper* are flag columns and hence scaling is not required"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# creating a copy to keep original df as is for later\n",
    "df_clust = df_features.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols_scale = [\n",
    "    'sales','units','upt','aur','aov','unique_categories_bought','unique_payments_used',\n",
    "    'unique_products_bought','orders']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1000, 26)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sales</th>\n",
       "      <th>units</th>\n",
       "      <th>orders</th>\n",
       "      <th>unique_products_bought</th>\n",
       "      <th>unique_payments_used</th>\n",
       "      <th>unique_categories_bought</th>\n",
       "      <th>aov</th>\n",
       "      <th>aur</th>\n",
       "      <th>upt</th>\n",
       "      <th>category_a_sales</th>\n",
       "      <th>category_b_sales</th>\n",
       "      <th>category_c_sales</th>\n",
       "      <th>category_d_sales</th>\n",
       "      <th>category_e_sales</th>\n",
       "      <th>category_a_units</th>\n",
       "      <th>category_b_units</th>\n",
       "      <th>category_c_units</th>\n",
       "      <th>category_d_units</th>\n",
       "      <th>category_e_units</th>\n",
       "      <th>payment_cash</th>\n",
       "      <th>payment_credit</th>\n",
       "      <th>payment_debit</th>\n",
       "      <th>payment_gc</th>\n",
       "      <th>payment_others</th>\n",
       "      <th>email_subscribed</th>\n",
       "      <th>omni_shopper</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.119330</td>\n",
       "      <td>0.145455</td>\n",
       "      <td>0.095238</td>\n",
       "      <td>0.1875</td>\n",
       "      <td>0.50</td>\n",
       "      <td>0.50</td>\n",
       "      <td>0.337757</td>\n",
       "      <td>0.371123</td>\n",
       "      <td>0.450000</td>\n",
       "      <td>0.480167</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.450939</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.068894</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.272727</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.519833</td>\n",
       "      <td>0.167015</td>\n",
       "      <td>0.313152</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.259174</td>\n",
       "      <td>0.218182</td>\n",
       "      <td>0.190476</td>\n",
       "      <td>0.2500</td>\n",
       "      <td>0.75</td>\n",
       "      <td>0.50</td>\n",
       "      <td>0.528931</td>\n",
       "      <td>0.614118</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>0.721703</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.247144</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.031153</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.031153</td>\n",
       "      <td>0.704050</td>\n",
       "      <td>0.186916</td>\n",
       "      <td>0.077882</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.228547</td>\n",
       "      <td>0.327273</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.4375</td>\n",
       "      <td>0.75</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.291533</td>\n",
       "      <td>0.338936</td>\n",
       "      <td>0.412500</td>\n",
       "      <td>0.240373</td>\n",
       "      <td>0.175029</td>\n",
       "      <td>0.309218</td>\n",
       "      <td>0.210035</td>\n",
       "      <td>0.065344</td>\n",
       "      <td>0.190476</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.095238</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.263711</td>\n",
       "      <td>0.309218</td>\n",
       "      <td>0.147025</td>\n",
       "      <td>0.280047</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.674372</td>\n",
       "      <td>0.745455</td>\n",
       "      <td>0.857143</td>\n",
       "      <td>0.8125</td>\n",
       "      <td>0.75</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.422360</td>\n",
       "      <td>0.500535</td>\n",
       "      <td>0.410714</td>\n",
       "      <td>0.575000</td>\n",
       "      <td>0.067500</td>\n",
       "      <td>0.158333</td>\n",
       "      <td>0.170833</td>\n",
       "      <td>0.028333</td>\n",
       "      <td>0.409091</td>\n",
       "      <td>0.159091</td>\n",
       "      <td>0.181818</td>\n",
       "      <td>0.113636</td>\n",
       "      <td>0.136364</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.345000</td>\n",
       "      <td>0.370000</td>\n",
       "      <td>0.218333</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.079168</td>\n",
       "      <td>0.090909</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0625</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.418002</td>\n",
       "      <td>0.358824</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.735294</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.735294</td>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                sales     units    orders  unique_products_bought  \\\n",
       "customer_id                                                         \n",
       "1            0.119330  0.145455  0.095238                  0.1875   \n",
       "2            0.259174  0.218182  0.190476                  0.2500   \n",
       "3            0.228547  0.327273  0.333333                  0.4375   \n",
       "4            0.674372  0.745455  0.857143                  0.8125   \n",
       "5            0.079168  0.090909  0.000000                  0.0625   \n",
       "\n",
       "             unique_payments_used  unique_categories_bought       aov  \\\n",
       "customer_id                                                             \n",
       "1                            0.50                      0.50  0.337757   \n",
       "2                            0.75                      0.50  0.528931   \n",
       "3                            0.75                      1.00  0.291533   \n",
       "4                            0.75                      1.00  0.422360   \n",
       "5                            0.25                      0.25  0.418002   \n",
       "\n",
       "                  aur       upt  category_a_sales  category_b_sales  \\\n",
       "customer_id                                                           \n",
       "1            0.371123  0.450000          0.480167          0.000000   \n",
       "2            0.614118  0.428571          0.721703          0.000000   \n",
       "3            0.338936  0.412500          0.240373          0.175029   \n",
       "4            0.500535  0.410714          0.575000          0.067500   \n",
       "5            0.358824  0.625000          0.000000          0.264706   \n",
       "\n",
       "             category_c_sales  category_d_sales  category_e_sales  \\\n",
       "customer_id                                                         \n",
       "1                    0.450939          0.000000          0.068894   \n",
       "2                    0.247144          0.000000          0.031153   \n",
       "3                    0.309218          0.210035          0.065344   \n",
       "4                    0.158333          0.170833          0.028333   \n",
       "5                    0.735294          0.000000          0.000000   \n",
       "\n",
       "             category_a_units  category_b_units  category_c_units  \\\n",
       "customer_id                                                         \n",
       "1                    0.363636          0.000000          0.363636   \n",
       "2                    0.466667          0.000000          0.333333   \n",
       "3                    0.190476          0.238095          0.238095   \n",
       "4                    0.409091          0.159091          0.181818   \n",
       "5                    0.000000          0.375000          0.625000   \n",
       "\n",
       "             category_d_units  category_e_units  payment_cash  payment_credit  \\\n",
       "customer_id                                                                     \n",
       "1                    0.000000          0.272727      0.000000        0.519833   \n",
       "2                    0.000000          0.200000      0.031153        0.704050   \n",
       "3                    0.095238          0.238095      0.000000        0.263711   \n",
       "4                    0.113636          0.136364      0.066667        0.345000   \n",
       "5                    0.000000          0.000000      0.000000        0.000000   \n",
       "\n",
       "             payment_debit  payment_gc  payment_others  email_subscribed  \\\n",
       "customer_id                                                                \n",
       "1                 0.167015    0.313152        0.000000                 0   \n",
       "2                 0.186916    0.077882        0.000000                 1   \n",
       "3                 0.309218    0.147025        0.280047                 0   \n",
       "4                 0.370000    0.218333        0.000000                 1   \n",
       "5                 0.735294    0.264706        0.000000                 0   \n",
       "\n",
       "             omni_shopper  \n",
       "customer_id                \n",
       "1                       0  \n",
       "2                       1  \n",
       "3                       0  \n",
       "4                       0  \n",
       "5                       0  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaler = MinMaxScaler().fit(df_clust[cols_scale])\n",
    "df_clust[cols_scale] = scaler.transform(df_clust[cols_scale])\n",
    "\n",
    "print(df_clust.shape)\n",
    "df_clust.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save the scaler object for scoring\n",
    "# save scaler\n",
    "pkl_filename = \"./files/model_objects/kmeans_scaler_model.pkl\"\n",
    "with open(pkl_filename, 'wb') as file:\n",
    "    pickle.dump(scaler, file)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Re-check feature summary</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>sales</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.310729</td>\n",
       "      <td>0.233232</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.144322</td>\n",
       "      <td>0.228258</td>\n",
       "      <td>0.400896</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>units</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.329236</td>\n",
       "      <td>0.261971</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.145455</td>\n",
       "      <td>0.236364</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>orders</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.324333</td>\n",
       "      <td>0.304047</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.095238</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_products_bought</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.346687</td>\n",
       "      <td>0.240658</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.187500</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.437500</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_payments_used</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.501250</td>\n",
       "      <td>0.237154</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique_categories_bought</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.710750</td>\n",
       "      <td>0.253646</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aov</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.425937</td>\n",
       "      <td>0.151708</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.327231</td>\n",
       "      <td>0.419629</td>\n",
       "      <td>0.512804</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>aur</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.496825</td>\n",
       "      <td>0.155156</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.394910</td>\n",
       "      <td>0.489831</td>\n",
       "      <td>0.591228</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>upt</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.421759</td>\n",
       "      <td>0.129003</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.337500</td>\n",
       "      <td>0.427354</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_a_sales</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.423622</td>\n",
       "      <td>0.232920</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.270173</td>\n",
       "      <td>0.428711</td>\n",
       "      <td>0.585021</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_b_sales</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.097276</td>\n",
       "      <td>0.110024</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.015163</td>\n",
       "      <td>0.074244</td>\n",
       "      <td>0.140781</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_c_sales</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.228259</td>\n",
       "      <td>0.185711</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.092650</td>\n",
       "      <td>0.205285</td>\n",
       "      <td>0.326688</td>\n",
       "      <td>0.951860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_d_sales</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.214959</td>\n",
       "      <td>0.206029</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.188518</td>\n",
       "      <td>0.350588</td>\n",
       "      <td>0.940171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_e_sales</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.035884</td>\n",
       "      <td>0.058451</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.022222</td>\n",
       "      <td>0.046804</td>\n",
       "      <td>0.809524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_a_units</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.320206</td>\n",
       "      <td>0.193179</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.187500</td>\n",
       "      <td>0.310728</td>\n",
       "      <td>0.437500</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_b_units</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.182344</td>\n",
       "      <td>0.156225</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.044960</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>0.277778</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_c_units</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.225336</td>\n",
       "      <td>0.170013</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.100000</td>\n",
       "      <td>0.207940</td>\n",
       "      <td>0.326629</td>\n",
       "      <td>0.857143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_d_units</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.135895</td>\n",
       "      <td>0.139798</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.106763</td>\n",
       "      <td>0.208576</td>\n",
       "      <td>0.727273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category_e_units</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.136219</td>\n",
       "      <td>0.144143</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.112374</td>\n",
       "      <td>0.208576</td>\n",
       "      <td>0.888889</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_cash</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.033727</td>\n",
       "      <td>0.085784</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.010809</td>\n",
       "      <td>0.947867</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_credit</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.542067</td>\n",
       "      <td>0.222836</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.403791</td>\n",
       "      <td>0.539482</td>\n",
       "      <td>0.693278</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_debit</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.268247</td>\n",
       "      <td>0.199722</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.123380</td>\n",
       "      <td>0.247137</td>\n",
       "      <td>0.383302</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_gc</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.128121</td>\n",
       "      <td>0.160347</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.072838</td>\n",
       "      <td>0.198798</td>\n",
       "      <td>0.968661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>payment_others</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.027838</td>\n",
       "      <td>0.070137</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.008834</td>\n",
       "      <td>0.605042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>email_subscribed</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.624000</td>\n",
       "      <td>0.484622</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>omni_shopper</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>0.249000</td>\n",
       "      <td>0.432650</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           count      mean       std  min       25%       50%  \\\n",
       "sales                     1000.0  0.310729  0.233232  0.0  0.144322  0.228258   \n",
       "units                     1000.0  0.329236  0.261971  0.0  0.145455  0.236364   \n",
       "orders                    1000.0  0.324333  0.304047  0.0  0.095238  0.238095   \n",
       "unique_products_bought    1000.0  0.346687  0.240658  0.0  0.187500  0.250000   \n",
       "unique_payments_used      1000.0  0.501250  0.237154  0.0  0.250000  0.500000   \n",
       "unique_categories_bought  1000.0  0.710750  0.253646  0.0  0.500000  0.750000   \n",
       "aov                       1000.0  0.425937  0.151708  0.0  0.327231  0.419629   \n",
       "aur                       1000.0  0.496825  0.155156  0.0  0.394910  0.489831   \n",
       "upt                       1000.0  0.421759  0.129003  0.0  0.337500  0.427354   \n",
       "category_a_sales          1000.0  0.423622  0.232920  0.0  0.270173  0.428711   \n",
       "category_b_sales          1000.0  0.097276  0.110024  0.0  0.015163  0.074244   \n",
       "category_c_sales          1000.0  0.228259  0.185711  0.0  0.092650  0.205285   \n",
       "category_d_sales          1000.0  0.214959  0.206029  0.0  0.000000  0.188518   \n",
       "category_e_sales          1000.0  0.035884  0.058451  0.0  0.000000  0.022222   \n",
       "category_a_units          1000.0  0.320206  0.193179  0.0  0.187500  0.310728   \n",
       "category_b_units          1000.0  0.182344  0.156225  0.0  0.044960  0.166667   \n",
       "category_c_units          1000.0  0.225336  0.170013  0.0  0.100000  0.207940   \n",
       "category_d_units          1000.0  0.135895  0.139798  0.0  0.000000  0.106763   \n",
       "category_e_units          1000.0  0.136219  0.144143  0.0  0.000000  0.112374   \n",
       "payment_cash              1000.0  0.033727  0.085784  0.0  0.000000  0.000000   \n",
       "payment_credit            1000.0  0.542067  0.222836  0.0  0.403791  0.539482   \n",
       "payment_debit             1000.0  0.268247  0.199722  0.0  0.123380  0.247137   \n",
       "payment_gc                1000.0  0.128121  0.160347  0.0  0.000000  0.072838   \n",
       "payment_others            1000.0  0.027838  0.070137  0.0  0.000000  0.000000   \n",
       "email_subscribed          1000.0  0.624000  0.484622  0.0  0.000000  1.000000   \n",
       "omni_shopper              1000.0  0.249000  0.432650  0.0  0.000000  0.000000   \n",
       "\n",
       "                               75%       max  \n",
       "sales                     0.400896  1.000000  \n",
       "units                     0.363636  1.000000  \n",
       "orders                    0.333333  1.000000  \n",
       "unique_products_bought    0.437500  1.000000  \n",
       "unique_payments_used      0.750000  1.000000  \n",
       "unique_categories_bought  1.000000  1.000000  \n",
       "aov                       0.512804  1.000000  \n",
       "aur                       0.591228  1.000000  \n",
       "upt                       0.500000  1.000000  \n",
       "category_a_sales          0.585021  1.000000  \n",
       "category_b_sales          0.140781  1.000000  \n",
       "category_c_sales          0.326688  0.951860  \n",
       "category_d_sales          0.350588  0.940171  \n",
       "category_e_sales          0.046804  0.809524  \n",
       "category_a_units          0.437500  1.000000  \n",
       "category_b_units          0.277778  1.000000  \n",
       "category_c_units          0.326629  0.857143  \n",
       "category_d_units          0.208576  0.727273  \n",
       "category_e_units          0.208576  0.888889  \n",
       "payment_cash              0.010809  0.947867  \n",
       "payment_credit            0.693278  1.000000  \n",
       "payment_debit             0.383302  1.000000  \n",
       "payment_gc                0.198798  0.968661  \n",
       "payment_others            0.008834  0.605042  \n",
       "email_subscribed          1.000000  1.000000  \n",
       "omni_shopper              0.000000  1.000000  "
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_clust.describe().T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Determining optimum k\n",
    "We will use elbow curve to determine optimum value of k for k-means clustering. The curve will be plotted for inertia score.\n",
    "\n",
    "We will also determine & plot silhouette score for all runs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████| 19/19 [00:03<00:00,  4.80it/s]\n"
     ]
    }
   ],
   "source": [
    "k1 = []\n",
    "sscore1 = []\n",
    "inertia_s1 = []\n",
    "\n",
    "for i in tqdm(range(2,21)):\n",
    "    k1.append(i)\n",
    "    kmeans1 = KMeans(n_clusters=i,random_state=125,max_iter=100).fit(df_clust)\n",
    "    # print('k-means done')\n",
    "    sscore1.append(ss(df_clust,kmeans1.labels_))\n",
    "    # print('ss score done')\n",
    "    inertia_s1.append(kmeans1.inertia_)\n",
    "    # print('inertia done')\n",
    "    # print('---')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Inertia score (SSE)')"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,5))\n",
    "plt.plot(k1,inertia_s1,color='blue', linestyle='dashed', marker='o',\n",
    "         markerfacecolor='red', markersize=10)\n",
    "plt.title('Inertia (SSE) vs. K Value')\n",
    "plt.xlabel('K')\n",
    "plt.ylabel('Inertia score (SSE)')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'SS')"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,5))\n",
    "plt.plot(k1,sscore1,color='blue', linestyle='dashed', marker='o',\n",
    "         markerfacecolor='red', markersize=10)\n",
    "plt.title('SS vs. K Value')\n",
    "plt.xlabel('K')\n",
    "plt.ylabel('SS')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Elbow is around 5-6. Using SS score, k can be either 5 or 6 as these have the highest SS scores</b>\n",
    "\n",
    "<b>Fixing k=6</b>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Final clustering iteration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sscore 0.2644425716291346\n",
      "inertia 591.0601380602852\n"
     ]
    }
   ],
   "source": [
    "kmeans = KMeans(n_clusters=6,random_state=125,max_iter=100).fit(df_clust)\n",
    "sscore = ss(df_clust,kmeans.labels_)\n",
    "inertia= kmeans.inertia_\n",
    "\n",
    "print('sscore',sscore)\n",
    "print('inertia',inertia)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save k-means\n",
    "pkl_filename = \"./files/model_objects/kmeans_model.pkl\"\n",
    "with open(pkl_filename, 'wb') as file:\n",
    "    pickle.dump(kmeans, file)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. Profiling\n",
    "\n",
    "We will use pandas .describe() to get cluster level summaries for profiling the clusters."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Add cluster labels</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features['cluster_ids'] = kmeans.labels_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Add *age* & *hh_income* from customer table to final features table</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_features = df_features.merge(\n",
    "    df_customer[['customer_id','age','hh_income']].set_index('customer_id'),\n",
    "    on='customer_id',how='left')\n",
    "df_features = df_features.reset_index()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Overall level summary</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_profile_overall = df_features.describe().T\n",
    "\n",
    "# use median for age and hh_income\n",
    "df_profile_overall['Overall Dataset'] = df_profile_overall.apply(\n",
    "    lambda row: row['50%']   if row.name in ['age','hh_income'] else\n",
    "    row['count'] if row.name in ['customer_id'] else row['mean'],\n",
    "    axis=1)\n",
    "df_profile_overall = df_profile_overall[['Overall Dataset']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Cluster ID level summary</b>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cluster_summary = df_features.groupby('cluster_ids').describe().T.reset_index()\n",
    "df_cluster_summary = df_cluster_summary.rename(columns={'level_0':'column','level_1':'metric'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# keep median for age & hh_income, count for customer id\n",
    "df_cluster_summary = df_cluster_summary.query(\n",
    "    '''\n",
    "        (column.isin([\"age\",\"hh_income\"]) & metric == \"50%\") | \\\n",
    "        (column.isin([\"customer_id\"]) & metric == \"count\") | \\\n",
    "        (~column.isin([\"customer_id\",\"age\",\"hh_income\"]) & metric == \"mean\")\n",
    "    ''',\n",
    "    engine='python')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_cluster_summary['Metric'] = df_cluster_summary.apply(\n",
    "    lambda row: 'median' if row['metric']=='50%' else row['metric'],\n",
    "    axis=1)\n",
    "df_cluster_summary = df_cluster_summary.set_index('column')\n",
    "df_cluster_summary = df_cluster_summary.drop('metric', axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Save profiling table as csv</b>\n",
    "\n",
    "We will then look at this table in excel to profile the clusters.\n",
    "- All clusters can be compared to one another and to the overall dataset across all features\n",
    "- Differences in features across clusters can be combined to describe each cluster\n",
    "- The data used above does not have any strong differentiating markers except email subscription and omni shopper flags\n",
    "- In real data you could have clusters such as *high value shoppers*, *low frequency shoppers*, or clusters showing affinity towards a single product category: *category B shoppers*, etc.\n",
    "\n",
    "<mark>Note</mark>: You could also go back and change the number of clusters in case the profiles aren't satisfactory. There are a few parameters as well that can be changed in the kMeans model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_profile = df_cluster_summary.join(df_profile_overall) # joins on Index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_profile.to_csv('./files/kmeans_profiling.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4. Scoring\n",
    "Once we have a k-means model trained and the clusters profiled, we can setup a scoring process that will assign cluster labels to new data.<br> This can help in tagging customers to particular clusters using latest transactions data and tracking cluster growth/movement over time.\n",
    "\n",
    "This part of the process most likely will be done using a python script that is scheduled to run at a particular cadence instead of being done via Jupyter. You will have a code or codes that prepare the data in the required format and then a python script that uses the saved scaling and model objects to score the updated data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Below process starts after the features dataset is created at customer id level\n",
    "# Please have required steps done prior to this step to have the features dataset ready\n",
    "\n",
    "# load features df\n",
    "df = pd.read_csv('./data/clustering_features.csv').set_index('customer_id')"
   ]
  },
  {
   "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>sales</th>\n",
       "      <th>units</th>\n",
       "      <th>orders</th>\n",
       "      <th>unique_products_bought</th>\n",
       "      <th>unique_payments_used</th>\n",
       "      <th>unique_categories_bought</th>\n",
       "      <th>aov</th>\n",
       "      <th>aur</th>\n",
       "      <th>upt</th>\n",
       "      <th>category_a_sales</th>\n",
       "      <th>category_b_sales</th>\n",
       "      <th>category_c_sales</th>\n",
       "      <th>category_d_sales</th>\n",
       "      <th>category_e_sales</th>\n",
       "      <th>category_a_units</th>\n",
       "      <th>category_b_units</th>\n",
       "      <th>category_c_units</th>\n",
       "      <th>category_d_units</th>\n",
       "      <th>category_e_units</th>\n",
       "      <th>payment_cash</th>\n",
       "      <th>payment_credit</th>\n",
       "      <th>payment_debit</th>\n",
       "      <th>payment_gc</th>\n",
       "      <th>payment_others</th>\n",
       "      <th>email_subscribed</th>\n",
       "      <th>omni_shopper</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2395</td>\n",
       "      <td>11</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>479.000000</td>\n",
       "      <td>217.727273</td>\n",
       "      <td>2.200000</td>\n",
       "      <td>0.480167</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.450939</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.068894</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.272727</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.519833</td>\n",
       "      <td>0.167015</td>\n",
       "      <td>0.313152</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4815</td>\n",
       "      <td>15</td>\n",
       "      <td>7</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>687.857143</td>\n",
       "      <td>321.000000</td>\n",
       "      <td>2.142857</td>\n",
       "      <td>0.721703</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.247144</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.031153</td>\n",
       "      <td>0.466667</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.031153</td>\n",
       "      <td>0.704050</td>\n",
       "      <td>0.186916</td>\n",
       "      <td>0.077882</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4285</td>\n",
       "      <td>21</td>\n",
       "      <td>10</td>\n",
       "      <td>9</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>428.500000</td>\n",
       "      <td>204.047619</td>\n",
       "      <td>2.100000</td>\n",
       "      <td>0.240373</td>\n",
       "      <td>0.175029</td>\n",
       "      <td>0.309218</td>\n",
       "      <td>0.210035</td>\n",
       "      <td>0.065344</td>\n",
       "      <td>0.190476</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>0.095238</td>\n",
       "      <td>0.238095</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.263711</td>\n",
       "      <td>0.309218</td>\n",
       "      <td>0.147025</td>\n",
       "      <td>0.280047</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12000</td>\n",
       "      <td>44</td>\n",
       "      <td>21</td>\n",
       "      <td>15</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>571.428571</td>\n",
       "      <td>272.727273</td>\n",
       "      <td>2.095238</td>\n",
       "      <td>0.575000</td>\n",
       "      <td>0.067500</td>\n",
       "      <td>0.158333</td>\n",
       "      <td>0.170833</td>\n",
       "      <td>0.028333</td>\n",
       "      <td>0.409091</td>\n",
       "      <td>0.159091</td>\n",
       "      <td>0.181818</td>\n",
       "      <td>0.113636</td>\n",
       "      <td>0.136364</td>\n",
       "      <td>0.066667</td>\n",
       "      <td>0.345000</td>\n",
       "      <td>0.370000</td>\n",
       "      <td>0.218333</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1700</td>\n",
       "      <td>8</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>566.666667</td>\n",
       "      <td>212.500000</td>\n",
       "      <td>2.666667</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.264706</td>\n",
       "      <td>0.735294</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.375000</td>\n",
       "      <td>0.625000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.735294</td>\n",
       "      <td>0.264706</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             sales  units  orders  unique_products_bought  \\\n",
       "customer_id                                                 \n",
       "1             2395     11       5                       5   \n",
       "2             4815     15       7                       6   \n",
       "3             4285     21      10                       9   \n",
       "4            12000     44      21                      15   \n",
       "5             1700      8       3                       3   \n",
       "\n",
       "             unique_payments_used  unique_categories_bought         aov  \\\n",
       "customer_id                                                               \n",
       "1                               3                         3  479.000000   \n",
       "2                               4                         3  687.857143   \n",
       "3                               4                         5  428.500000   \n",
       "4                               4                         5  571.428571   \n",
       "5                               2                         2  566.666667   \n",
       "\n",
       "                    aur       upt  category_a_sales  category_b_sales  \\\n",
       "customer_id                                                             \n",
       "1            217.727273  2.200000          0.480167               NaN   \n",
       "2            321.000000  2.142857          0.721703               NaN   \n",
       "3            204.047619  2.100000          0.240373          0.175029   \n",
       "4            272.727273  2.095238          0.575000          0.067500   \n",
       "5            212.500000  2.666667               NaN          0.264706   \n",
       "\n",
       "             category_c_sales  category_d_sales  category_e_sales  \\\n",
       "customer_id                                                         \n",
       "1                    0.450939               NaN          0.068894   \n",
       "2                    0.247144               NaN          0.031153   \n",
       "3                    0.309218          0.210035          0.065344   \n",
       "4                    0.158333          0.170833          0.028333   \n",
       "5                    0.735294               NaN               NaN   \n",
       "\n",
       "             category_a_units  category_b_units  category_c_units  \\\n",
       "customer_id                                                         \n",
       "1                    0.363636               NaN          0.363636   \n",
       "2                    0.466667               NaN          0.333333   \n",
       "3                    0.190476          0.238095          0.238095   \n",
       "4                    0.409091          0.159091          0.181818   \n",
       "5                         NaN          0.375000          0.625000   \n",
       "\n",
       "             category_d_units  category_e_units  payment_cash  payment_credit  \\\n",
       "customer_id                                                                     \n",
       "1                         NaN          0.272727           NaN        0.519833   \n",
       "2                         NaN          0.200000      0.031153        0.704050   \n",
       "3                    0.095238          0.238095           NaN        0.263711   \n",
       "4                    0.113636          0.136364      0.066667        0.345000   \n",
       "5                         NaN               NaN           NaN             NaN   \n",
       "\n",
       "             payment_debit  payment_gc  payment_others  email_subscribed  \\\n",
       "customer_id                                                                \n",
       "1                 0.167015    0.313152             NaN                 0   \n",
       "2                 0.186916    0.077882             NaN                 1   \n",
       "3                 0.309218    0.147025        0.280047                 0   \n",
       "4                 0.370000    0.218333             NaN                 1   \n",
       "5                 0.735294    0.264706             NaN                 0   \n",
       "\n",
       "             omni_shopper  \n",
       "customer_id                \n",
       "1                       0  \n",
       "2                       1  \n",
       "3                       0  \n",
       "4                       0  \n",
       "5                       0  "
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "# fill na\n",
    "df = df.fillna(0)\n",
    "\n",
    "# scaling\n",
    "cols_scale = [\n",
    "    'sales','units','upt','aur','aov','unique_categories_bought','unique_payments_used',\n",
    "    'unique_products_bought','orders']\n",
    "\n",
    "# Load scaler  from file\n",
    "pkl_filename = \"./files/model_objects/kmeans_scaler_model.pkl\"\n",
    "with open(pkl_filename, 'rb') as file:\n",
    "    scaler = pickle.load(file)\n",
    "\n",
    "df[cols_scale] = scaler.transform(df[cols_scale])\n",
    "\n",
    "# Load k-means from file\n",
    "pkl_filename = \"./files/model_objects/kmeans_model.pkl\"\n",
    "with open(pkl_filename, 'rb') as file:\n",
    "    kmeans = pickle.load(file)\n",
    "\n",
    "df['cluster_ids'] = kmeans.predict(df)\n",
    "\n",
    "# save labels\n",
    "df[['cluster_ids']].to_csv('./files/kmeans_labels.csv', index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    293\n",
       "4    284\n",
       "2    194\n",
       "1     92\n",
       "5     82\n",
       "3     55\n",
       "Name: cluster_ids, dtype: int64"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.cluster_ids.value_counts()"
   ]
  },
  {
   "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}