buy-till-you-die-models-customer-lifetime-value.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.015617,
     "end_time": "2021-02-06T11:59:01.877099",
     "exception": false,
     "start_time": "2021-02-06T11:59:01.861482",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# **Buy Till You Die Models: Customer Lifetime Value**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.014753,
     "end_time": "2021-02-06T11:59:01.906549",
     "exception": false,
     "start_time": "2021-02-06T11:59:01.891796",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "<p><img style=\"float: left;margin:20px 10px 5px 1px; max-width:380px\" src=\"https://i.ytimg.com/vi/l_xwYM2KNX0/maxresdefault.jpg\"></p>\n",
    "<p> The Buy Till You Die (BTYD) class of statistical models is designed to predict the behavioral characteristics of customers. The main purpose is to model and forecast customer lifetime value. </p>\n",
    "<p> Customer lifetime value (CLTV or CLV) can be defined as the present value of a customer for the company based on projected future cash flows from the customer relationship. CLTV represents the total amount of money spent on the business or products over lifetime of a customer.</p>\n",
    "<p> In this notebook, I'll try to apply a BTYD model on an online retail dataset to make a 1-year CLTV prediction.</p> "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.013992,
     "end_time": "2021-02-06T11:59:01.934814",
     "exception": false,
     "start_time": "2021-02-06T11:59:01.920822",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## 1. Loading the Dataset & Checking Variables"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.014215,
     "end_time": "2021-02-06T11:59:01.963315",
     "exception": false,
     "start_time": "2021-02-06T11:59:01.949100",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "\n",
    "### ****Content of Dataset****\n",
    "\n",
    "This Online Retail II data set contains all the transactions occurring for a UK-based and registered, non-store online retail between 01/12/2009 and 09/12/2011.The company mainly sells unique all-occasion gift-ware. Many customers of the company are wholesalers."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.013987,
     "end_time": "2021-02-06T11:59:01.991827",
     "exception": false,
     "start_time": "2021-02-06T11:59:01.977840",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### ****Definiton of Variables****\n",
    "1. **Invoice:** Invoice number, unique identifier variable for each transaction. Refund invoice numbers starts with \"C\"\n",
    "2. **StockCode:** Unique product code\n",
    "3. **Description:** Product name\n",
    "4. **Quantity:** The number of product in the invoice\n",
    "5. **InvoiceDate:** Date and time of the purchase\n",
    "6. **Price:** Unit price of a product (in terms of Sterlin)\n",
    "7. **CustomerID:** Unique customer identifier\n",
    "8. **Country:** Residential country of customers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_kg_hide-output": true,
    "collapsed": true,
    "execution": {
     "iopub.execute_input": "2021-02-06T11:59:02.025440Z",
     "iopub.status.busy": "2021-02-06T11:59:02.024743Z",
     "iopub.status.idle": "2021-02-06T11:59:12.074131Z",
     "shell.execute_reply": "2021-02-06T11:59:12.073396Z"
    },
    "papermill": {
     "duration": 10.066683,
     "end_time": "2021-02-06T11:59:12.074356",
     "exception": false,
     "start_time": "2021-02-06T11:59:02.007673",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting openpyxl\r\n",
      "  Downloading openpyxl-3.0.6-py2.py3-none-any.whl (242 kB)\r\n",
      "\u001b[K     |████████████████████████████████| 242 kB 893 kB/s \r\n",
      "\u001b[?25hCollecting et-xmlfile\r\n",
      "  Downloading et_xmlfile-1.0.1.tar.gz (8.4 kB)\r\n",
      "Collecting jdcal\r\n",
      "  Downloading jdcal-1.4.1-py2.py3-none-any.whl (9.5 kB)\r\n",
      "Building wheels for collected packages: et-xmlfile\r\n",
      "  Building wheel for et-xmlfile (setup.py) ... \u001b[?25l-\b \b\\\b \bdone\r\n",
      "\u001b[?25h  Created wheel for et-xmlfile: filename=et_xmlfile-1.0.1-py3-none-any.whl size=8913 sha256=7b3a4d6d68ca0b9e4d70a9cef1d712215c981222da1479b5ee898eda7ed2a0f3\r\n",
      "  Stored in directory: /root/.cache/pip/wheels/e2/bd/55/048b4fd505716c4c298f42ee02dffd9496bb6d212b266c7f31\r\n",
      "Successfully built et-xmlfile\r\n",
      "Installing collected packages: jdcal, et-xmlfile, openpyxl\r\n",
      "Successfully installed et-xmlfile-1.0.1 jdcal-1.4.1 openpyxl-3.0.6\r\n",
      "\u001b[33mWARNING: You are using pip version 21.0; however, version 21.0.1 is available.\r\n",
      "You should consider upgrading via the '/opt/conda/bin/python -m pip install --upgrade pip' command.\u001b[0m\r\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install openpyxl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "_kg_hide-output": true,
    "collapsed": true,
    "execution": {
     "iopub.execute_input": "2021-02-06T11:59:12.121133Z",
     "iopub.status.busy": "2021-02-06T11:59:12.117655Z",
     "iopub.status.idle": "2021-02-06T11:59:18.529889Z",
     "shell.execute_reply": "2021-02-06T11:59:18.528999Z"
    },
    "papermill": {
     "duration": 6.436094,
     "end_time": "2021-02-06T11:59:18.530106",
     "exception": false,
     "start_time": "2021-02-06T11:59:12.094012",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting xlrd\r\n",
      "  Downloading xlrd-2.0.1-py2.py3-none-any.whl (96 kB)\r\n",
      "\u001b[K     |████████████████████████████████| 96 kB 726 kB/s \r\n",
      "\u001b[?25hInstalling collected packages: xlrd\r\n",
      "Successfully installed xlrd-2.0.1\r\n",
      "\u001b[33mWARNING: You are using pip version 21.0; however, version 21.0.1 is available.\r\n",
      "You should consider upgrading via the '/opt/conda/bin/python -m pip install --upgrade pip' command.\u001b[0m\r\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install xlrd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T11:59:18.582905Z",
     "iopub.status.busy": "2021-02-06T11:59:18.582336Z",
     "iopub.status.idle": "2021-02-06T11:59:18.585233Z",
     "shell.execute_reply": "2021-02-06T11:59:18.584808Z"
    },
    "papermill": {
     "duration": 0.028645,
     "end_time": "2021-02-06T11:59:18.585384",
     "exception": false,
     "start_time": "2021-02-06T11:59:18.556739",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Loading the required libraries\n",
    "import datetime as dt\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T11:59:18.634507Z",
     "iopub.status.busy": "2021-02-06T11:59:18.633910Z",
     "iopub.status.idle": "2021-02-06T12:00:38.005014Z",
     "shell.execute_reply": "2021-02-06T12:00:38.005503Z"
    },
    "papermill": {
     "duration": 79.397932,
     "end_time": "2021-02-06T12:00:38.005682",
     "exception": false,
     "start_time": "2021-02-06T11:59:18.607750",
     "status": "completed"
    },
    "tags": []
   },
   "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>Invoice</th>\n",
       "      <th>StockCode</th>\n",
       "      <th>Description</th>\n",
       "      <th>Quantity</th>\n",
       "      <th>InvoiceDate</th>\n",
       "      <th>Price</th>\n",
       "      <th>Customer ID</th>\n",
       "      <th>Country</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>536365</td>\n",
       "      <td>85123A</td>\n",
       "      <td>WHITE HANGING HEART T-LIGHT HOLDER</td>\n",
       "      <td>6</td>\n",
       "      <td>2010-12-01 08:26:00</td>\n",
       "      <td>2.55</td>\n",
       "      <td>17850.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>536365</td>\n",
       "      <td>71053</td>\n",
       "      <td>WHITE METAL LANTERN</td>\n",
       "      <td>6</td>\n",
       "      <td>2010-12-01 08:26:00</td>\n",
       "      <td>3.39</td>\n",
       "      <td>17850.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>536365</td>\n",
       "      <td>84406B</td>\n",
       "      <td>CREAM CUPID HEARTS COAT HANGER</td>\n",
       "      <td>8</td>\n",
       "      <td>2010-12-01 08:26:00</td>\n",
       "      <td>2.75</td>\n",
       "      <td>17850.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>536365</td>\n",
       "      <td>84029G</td>\n",
       "      <td>KNITTED UNION FLAG HOT WATER BOTTLE</td>\n",
       "      <td>6</td>\n",
       "      <td>2010-12-01 08:26:00</td>\n",
       "      <td>3.39</td>\n",
       "      <td>17850.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>536365</td>\n",
       "      <td>84029E</td>\n",
       "      <td>RED WOOLLY HOTTIE WHITE HEART.</td>\n",
       "      <td>6</td>\n",
       "      <td>2010-12-01 08:26:00</td>\n",
       "      <td>3.39</td>\n",
       "      <td>17850.0</td>\n",
       "      <td>United Kingdom</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Invoice StockCode                          Description  Quantity  \\\n",
       "0  536365    85123A   WHITE HANGING HEART T-LIGHT HOLDER         6   \n",
       "1  536365     71053                  WHITE METAL LANTERN         6   \n",
       "2  536365    84406B       CREAM CUPID HEARTS COAT HANGER         8   \n",
       "3  536365    84029G  KNITTED UNION FLAG HOT WATER BOTTLE         6   \n",
       "4  536365    84029E       RED WOOLLY HOTTIE WHITE HEART.         6   \n",
       "\n",
       "          InvoiceDate  Price  Customer ID         Country  \n",
       "0 2010-12-01 08:26:00   2.55      17850.0  United Kingdom  \n",
       "1 2010-12-01 08:26:00   3.39      17850.0  United Kingdom  \n",
       "2 2010-12-01 08:26:00   2.75      17850.0  United Kingdom  \n",
       "3 2010-12-01 08:26:00   3.39      17850.0  United Kingdom  \n",
       "4 2010-12-01 08:26:00   3.39      17850.0  United Kingdom  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Reading the online retail dataset\n",
    "df_ = pd.read_excel(\"../input/online-retail-ii-data-set-from-ml-repository/online_retail_II.xlsx\",sheet_name=\"Year 2010-2011\")\n",
    "\n",
    "# Copying the online retail dataset \n",
    "df = df_.copy()\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:38.063497Z",
     "iopub.status.busy": "2021-02-06T12:00:38.062562Z",
     "iopub.status.idle": "2021-02-06T12:00:38.132503Z",
     "shell.execute_reply": "2021-02-06T12:00:38.131631Z"
    },
    "papermill": {
     "duration": 0.105552,
     "end_time": "2021-02-06T12:00:38.132652",
     "exception": false,
     "start_time": "2021-02-06T12:00:38.027100",
     "status": "completed"
    },
    "scrolled": true,
    "tags": []
   },
   "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>Quantity</th>\n",
       "      <td>541910.0</td>\n",
       "      <td>9.552234</td>\n",
       "      <td>218.080957</td>\n",
       "      <td>-80995.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>80995.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Price</th>\n",
       "      <td>541910.0</td>\n",
       "      <td>4.611138</td>\n",
       "      <td>96.759765</td>\n",
       "      <td>-11062.06</td>\n",
       "      <td>1.25</td>\n",
       "      <td>2.08</td>\n",
       "      <td>4.13</td>\n",
       "      <td>38970.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <td>406830.0</td>\n",
       "      <td>15287.684160</td>\n",
       "      <td>1713.603074</td>\n",
       "      <td>12346.00</td>\n",
       "      <td>13953.00</td>\n",
       "      <td>15152.00</td>\n",
       "      <td>16791.00</td>\n",
       "      <td>18287.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                count          mean          std       min       25%  \\\n",
       "Quantity     541910.0      9.552234   218.080957 -80995.00      1.00   \n",
       "Price        541910.0      4.611138    96.759765 -11062.06      1.25   \n",
       "Customer ID  406830.0  15287.684160  1713.603074  12346.00  13953.00   \n",
       "\n",
       "                  50%       75%      max  \n",
       "Quantity         3.00     10.00  80995.0  \n",
       "Price            2.08      4.13  38970.0  \n",
       "Customer ID  15152.00  16791.00  18287.0  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking numerical variables\n",
    "df.describe().T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.022033,
     "end_time": "2021-02-06T12:00:38.176738",
     "exception": false,
     "start_time": "2021-02-06T12:00:38.154705",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "<p> When we look at the descriptive statistics of numerical variables, we can see that there are negative values in the Quantity and Price columns. These transactions are canceled orders. In the next step (data preprocessing), we'll eliminate these observations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:38.224829Z",
     "iopub.status.busy": "2021-02-06T12:00:38.224190Z",
     "iopub.status.idle": "2021-02-06T12:00:38.335432Z",
     "shell.execute_reply": "2021-02-06T12:00:38.334890Z"
    },
    "papermill": {
     "duration": 0.136762,
     "end_time": "2021-02-06T12:00:38.335573",
     "exception": false,
     "start_time": "2021-02-06T12:00:38.198811",
     "status": "completed"
    },
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Invoice             0\n",
       "StockCode           0\n",
       "Description      1454\n",
       "Quantity            0\n",
       "InvoiceDate         0\n",
       "Price               0\n",
       "Customer ID    135080\n",
       "Country             0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking null variables \n",
    "df.isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.021909,
     "end_time": "2021-02-06T12:00:38.379686",
     "exception": false,
     "start_time": "2021-02-06T12:00:38.357777",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "There are null observations in the dataset, especially in the Customer ID column. Since this analysis will be consumer-based, we need to remove these observations from the dataset. Therefore, we will  eliminate these observations in the next step (data preprocessing)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.022449,
     "end_time": "2021-02-06T12:00:38.424631",
     "exception": false,
     "start_time": "2021-02-06T12:00:38.402182",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## 2. Data Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:38.475771Z",
     "iopub.status.busy": "2021-02-06T12:00:38.475133Z",
     "iopub.status.idle": "2021-02-06T12:00:39.216371Z",
     "shell.execute_reply": "2021-02-06T12:00:39.215885Z"
    },
    "papermill": {
     "duration": 0.76946,
     "end_time": "2021-02-06T12:00:39.216526",
     "exception": false,
     "start_time": "2021-02-06T12:00:38.447066",
     "status": "completed"
    },
    "scrolled": true,
    "tags": []
   },
   "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>1%</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>99%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Quantity</th>\n",
       "      <td>397925.0</td>\n",
       "      <td>13.021793</td>\n",
       "      <td>180.419984</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>120.00</td>\n",
       "      <td>80995.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Price</th>\n",
       "      <td>397925.0</td>\n",
       "      <td>3.116212</td>\n",
       "      <td>22.096773</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.21</td>\n",
       "      <td>1.25</td>\n",
       "      <td>1.95</td>\n",
       "      <td>3.75</td>\n",
       "      <td>14.95</td>\n",
       "      <td>8142.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <td>397925.0</td>\n",
       "      <td>15294.308601</td>\n",
       "      <td>1713.172738</td>\n",
       "      <td>12346.0</td>\n",
       "      <td>12415.00</td>\n",
       "      <td>13969.00</td>\n",
       "      <td>15159.00</td>\n",
       "      <td>16795.00</td>\n",
       "      <td>18211.00</td>\n",
       "      <td>18287.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                count          mean          std      min        1%       25%  \\\n",
       "Quantity     397925.0     13.021793   180.419984      1.0      1.00      2.00   \n",
       "Price        397925.0      3.116212    22.096773      0.0      0.21      1.25   \n",
       "Customer ID  397925.0  15294.308601  1713.172738  12346.0  12415.00  13969.00   \n",
       "\n",
       "                  50%       75%       99%       max  \n",
       "Quantity         6.00     12.00    120.00  80995.00  \n",
       "Price            1.95      3.75     14.95   8142.75  \n",
       "Customer ID  15159.00  16795.00  18211.00  18287.00  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Data preparation step 1: Removing null oberservations\n",
    "df.dropna(inplace=True)\n",
    "\n",
    "# Data preparation step 1: Removing canceled orders \n",
    "df = df[~df[\"Invoice\"].str.contains(\"C\", na=False)]\n",
    "df = df[df[\"Quantity\"] > 0]\n",
    "\n",
    "df.describe([0.01,0.25,0.50,0.75,0.99]).T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.023181,
     "end_time": "2021-02-06T12:00:39.262727",
     "exception": false,
     "start_time": "2021-02-06T12:00:39.239546",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "<p>Check out descriptive statistics of numerical variables again! Look at 99% quantile of Quantity and Price columns and compare with the maximum values. We can say that there are some outliers.</p>\n",
    "<p>Let's find out these outliers and replace them with the highest limit.</p>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:39.317265Z",
     "iopub.status.busy": "2021-02-06T12:00:39.316619Z",
     "iopub.status.idle": "2021-02-06T12:00:39.318668Z",
     "shell.execute_reply": "2021-02-06T12:00:39.319095Z"
    },
    "papermill": {
     "duration": 0.032924,
     "end_time": "2021-02-06T12:00:39.319287",
     "exception": false,
     "start_time": "2021-02-06T12:00:39.286363",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Defining functions for outliers\n",
    "def outlier_thresholds(dataframe, variable):\n",
    "    quartile1 = dataframe[variable].quantile(0.01)\n",
    "    quartile3 = dataframe[variable].quantile(0.99)\n",
    "    interquantile_range = quartile3 - quartile1\n",
    "    up_limit = quartile3 + 1.5 * interquantile_range\n",
    "    low_limit = quartile1 - 1.5 * interquantile_range\n",
    "    return low_limit, up_limit\n",
    "\n",
    "# Defining functions to replace outliers\n",
    "def replace_with_thresholds(dataframe, variable):\n",
    "    low_limit, up_limit = outlier_thresholds(dataframe, variable)\n",
    "    # dataframe.loc[(dataframe[variable] < low_limit), variable] = low_limit\n",
    "    dataframe.loc[(dataframe[variable] > up_limit), variable] = up_limit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:39.369333Z",
     "iopub.status.busy": "2021-02-06T12:00:39.368731Z",
     "iopub.status.idle": "2021-02-06T12:00:39.474194Z",
     "shell.execute_reply": "2021-02-06T12:00:39.473641Z"
    },
    "papermill": {
     "duration": 0.131812,
     "end_time": "2021-02-06T12:00:39.474353",
     "exception": false,
     "start_time": "2021-02-06T12:00:39.342541",
     "status": "completed"
    },
    "tags": []
   },
   "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>1%</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>99%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Quantity</th>\n",
       "      <td>397925.0</td>\n",
       "      <td>11.833709</td>\n",
       "      <td>25.534486</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>120.00</td>\n",
       "      <td>298.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Price</th>\n",
       "      <td>397925.0</td>\n",
       "      <td>2.893201</td>\n",
       "      <td>3.227143</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.21</td>\n",
       "      <td>1.25</td>\n",
       "      <td>1.95</td>\n",
       "      <td>3.75</td>\n",
       "      <td>14.95</td>\n",
       "      <td>37.06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <td>397925.0</td>\n",
       "      <td>15294.308601</td>\n",
       "      <td>1713.172738</td>\n",
       "      <td>12346.0</td>\n",
       "      <td>12415.00</td>\n",
       "      <td>13969.00</td>\n",
       "      <td>15159.00</td>\n",
       "      <td>16795.00</td>\n",
       "      <td>18211.00</td>\n",
       "      <td>18287.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                count          mean          std      min        1%       25%  \\\n",
       "Quantity     397925.0     11.833709    25.534486      1.0      1.00      2.00   \n",
       "Price        397925.0      2.893201     3.227143      0.0      0.21      1.25   \n",
       "Customer ID  397925.0  15294.308601  1713.172738  12346.0  12415.00  13969.00   \n",
       "\n",
       "                  50%       75%       99%       max  \n",
       "Quantity         6.00     12.00    120.00    298.50  \n",
       "Price            1.95      3.75     14.95     37.06  \n",
       "Customer ID  15159.00  16795.00  18211.00  18287.00  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Data preparation step 2: Replacing outliers in the Quantity and Price columns with the upper limit\n",
    "replace_with_thresholds(df, \"Quantity\")\n",
    "replace_with_thresholds(df, \"Price\")\n",
    "\n",
    "df.describe([0.01,0.25,0.50,0.75,0.99]).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:39.527594Z",
     "iopub.status.busy": "2021-02-06T12:00:39.526933Z",
     "iopub.status.idle": "2021-02-06T12:00:39.532508Z",
     "shell.execute_reply": "2021-02-06T12:00:39.531872Z"
    },
    "papermill": {
     "duration": 0.034684,
     "end_time": "2021-02-06T12:00:39.532647",
     "exception": false,
     "start_time": "2021-02-06T12:00:39.497963",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Data preparation step 3: Calculating total price per transaction \n",
    "df[\"TotalPrice\"] = df[\"Quantity\"] * df[\"Price\"]\n",
    "today_date = dt.datetime(2011, 12, 11)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:39.585162Z",
     "iopub.status.busy": "2021-02-06T12:00:39.584603Z",
     "iopub.status.idle": "2021-02-06T12:00:39.590329Z",
     "shell.execute_reply": "2021-02-06T12:00:39.590745Z"
    },
    "papermill": {
     "duration": 0.034512,
     "end_time": "2021-02-06T12:00:39.590915",
     "exception": false,
     "start_time": "2021-02-06T12:00:39.556403",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Maximum invoice date: 2011-12-09 12:50:00 \n",
      " Today date: 2011-12-11 00:00:00\n"
     ]
    }
   ],
   "source": [
    "# Defining today date as max(InvoiceDate) + 2 days\n",
    "today_date = dt.datetime(2011, 12, 11)\n",
    "print(f\" Maximum invoice date: {df.InvoiceDate.max()} \\n Today date: {today_date}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.024285,
     "end_time": "2021-02-06T12:00:39.639927",
     "exception": false,
     "start_time": "2021-02-06T12:00:39.615642",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## 3. Deriving the RFM Metrics\n",
    "\n",
    "<p><img style=\"float: right;margin:-10px 20px 20px 5px; max-width:380px\" src=\"https://www.retailreco.com/blog/wp-content/uploads/2018/11/RFM-Analytics.jpg\"></p>\n",
    "\n",
    "The predictive CLTV models are built around 4 key metrics. These are:\n",
    "\n",
    "**<p>Recency:**  The age of the customer at the time of their last purchase.\n",
    "**<p>Monetary:** The average total sales of the customer.\n",
    "**<p>Frequency:** Number of purchases/transactions.\n",
    "**<p>Age (T):** The age of the customer since the date of a customer's first purchase to the current date."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:39.696145Z",
     "iopub.status.busy": "2021-02-06T12:00:39.695490Z",
     "iopub.status.idle": "2021-02-06T12:00:41.955151Z",
     "shell.execute_reply": "2021-02-06T12:00:41.955667Z"
    },
    "papermill": {
     "duration": 2.290952,
     "end_time": "2021-02-06T12:00:41.955844",
     "exception": false,
     "start_time": "2021-02-06T12:00:39.664892",
     "status": "completed"
    },
    "tags": []
   },
   "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>Recency</th>\n",
       "      <th>T</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Monetary</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12346.0</th>\n",
       "      <td>0</td>\n",
       "      <td>326</td>\n",
       "      <td>1</td>\n",
       "      <td>310.440000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12347.0</th>\n",
       "      <td>365</td>\n",
       "      <td>368</td>\n",
       "      <td>7</td>\n",
       "      <td>615.714286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12348.0</th>\n",
       "      <td>282</td>\n",
       "      <td>359</td>\n",
       "      <td>4</td>\n",
       "      <td>442.695000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12349.0</th>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>1</td>\n",
       "      <td>1491.720000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12350.0</th>\n",
       "      <td>0</td>\n",
       "      <td>311</td>\n",
       "      <td>1</td>\n",
       "      <td>331.460000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Recency    T  Frequency     Monetary\n",
       "Customer ID                                      \n",
       "12346.0            0  326          1   310.440000\n",
       "12347.0          365  368          7   615.714286\n",
       "12348.0          282  359          4   442.695000\n",
       "12349.0            0   19          1  1491.720000\n",
       "12350.0            0  311          1   331.460000"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculating recency, monetary, frequency and tenure metrics\n",
    "rfm = df.groupby(\"Customer ID\").agg({\"InvoiceDate\": [lambda date: (date.max() - date.min()).days,\n",
    "                                                     lambda date: (today_date - date.min()).days],\n",
    "                                     \"Invoice\": lambda num: num.nunique(),\n",
    "                                      \"TotalPrice\": lambda price: price.sum()}) #total price per customer\n",
    "\n",
    "rfm.columns = rfm.columns.droplevel(0)\n",
    "rfm.columns = ['Recency', \"T\", 'Frequency', \"Monetary\"]\n",
    "\n",
    "# Calculating average monetary values per order:\n",
    "rfm[\"Monetary\"] = rfm[\"Monetary\"] / rfm[\"Frequency\"]\n",
    "\n",
    "rfm.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:42.011501Z",
     "iopub.status.busy": "2021-02-06T12:00:42.010459Z",
     "iopub.status.idle": "2021-02-06T12:00:42.022617Z",
     "shell.execute_reply": "2021-02-06T12:00:42.021927Z"
    },
    "papermill": {
     "duration": 0.042359,
     "end_time": "2021-02-06T12:00:42.022786",
     "exception": false,
     "start_time": "2021-02-06T12:00:41.980427",
     "status": "completed"
    },
    "tags": []
   },
   "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>Recency</th>\n",
       "      <th>T</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Monetary</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Customer ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12347.0</th>\n",
       "      <td>365</td>\n",
       "      <td>368</td>\n",
       "      <td>7</td>\n",
       "      <td>615.714286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12348.0</th>\n",
       "      <td>282</td>\n",
       "      <td>359</td>\n",
       "      <td>4</td>\n",
       "      <td>442.695000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12352.0</th>\n",
       "      <td>260</td>\n",
       "      <td>297</td>\n",
       "      <td>8</td>\n",
       "      <td>219.542500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12356.0</th>\n",
       "      <td>302</td>\n",
       "      <td>326</td>\n",
       "      <td>3</td>\n",
       "      <td>937.143333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12358.0</th>\n",
       "      <td>149</td>\n",
       "      <td>151</td>\n",
       "      <td>2</td>\n",
       "      <td>575.210000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Recency    T  Frequency    Monetary\n",
       "Customer ID                                     \n",
       "12347.0          365  368          7  615.714286\n",
       "12348.0          282  359          4  442.695000\n",
       "12352.0          260  297          8  219.542500\n",
       "12356.0          302  326          3  937.143333\n",
       "12358.0          149  151          2  575.210000"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Removing one-time purchases from dataset\n",
    "rfm = rfm[(rfm['Frequency'] > 1)]\n",
    "\n",
    "# Copying dataset\n",
    "cltv = rfm.copy()\n",
    "rfm.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.025133,
     "end_time": "2021-02-06T12:00:42.073288",
     "exception": false,
     "start_time": "2021-02-06T12:00:42.048155",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# 4. Train BG-NBD Model (Expected number of transaction)\n",
    "\n",
    "In this step, we will train the BG/NBD model to estimate the expected sales in a given period of time.\n",
    "Also, the trained BG/NBD model will be used in predicting customer lifetime value in the following steps. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "_kg_hide-output": true,
    "collapsed": true,
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:42.135934Z",
     "iopub.status.busy": "2021-02-06T12:00:42.128523Z",
     "iopub.status.idle": "2021-02-06T12:00:48.831682Z",
     "shell.execute_reply": "2021-02-06T12:00:48.832663Z"
    },
    "papermill": {
     "duration": 6.733913,
     "end_time": "2021-02-06T12:00:48.832893",
     "exception": false,
     "start_time": "2021-02-06T12:00:42.098980",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting lifetimes\r\n",
      "  Downloading Lifetimes-0.11.3-py3-none-any.whl (584 kB)\r\n",
      "\u001b[K     |████████████████████████████████| 584 kB 889 kB/s \r\n",
      "\u001b[?25hRequirement already satisfied: autograd>=1.2.0 in /opt/conda/lib/python3.7/site-packages (from lifetimes) (1.3)\r\n",
      "Requirement already satisfied: pandas>=0.24.0 in /opt/conda/lib/python3.7/site-packages (from lifetimes) (1.2.0)\r\n",
      "Requirement already satisfied: dill>=0.2.6 in /opt/conda/lib/python3.7/site-packages (from lifetimes) (0.3.3)\r\n",
      "Requirement already satisfied: scipy>=1.0.0 in /opt/conda/lib/python3.7/site-packages (from lifetimes) (1.4.1)\r\n",
      "Requirement already satisfied: numpy>=1.10.0 in /opt/conda/lib/python3.7/site-packages (from lifetimes) (1.19.5)\r\n",
      "Requirement already satisfied: future>=0.15.2 in /opt/conda/lib/python3.7/site-packages (from autograd>=1.2.0->lifetimes) (0.18.2)\r\n",
      "Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/lib/python3.7/site-packages (from pandas>=0.24.0->lifetimes) (2.8.1)\r\n",
      "Requirement already satisfied: pytz>=2017.3 in /opt/conda/lib/python3.7/site-packages (from pandas>=0.24.0->lifetimes) (2019.3)\r\n",
      "Requirement already satisfied: six>=1.5 in /opt/conda/lib/python3.7/site-packages (from python-dateutil>=2.7.3->pandas>=0.24.0->lifetimes) (1.15.0)\r\n",
      "Installing collected packages: lifetimes\r\n",
      "Successfully installed lifetimes-0.11.3\r\n",
      "\u001b[33mWARNING: You are using pip version 21.0; however, version 21.0.1 is available.\r\n",
      "You should consider upgrading via the '/opt/conda/bin/python -m pip install --upgrade pip' command.\u001b[0m\r\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install lifetimes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:48.897308Z",
     "iopub.status.busy": "2021-02-06T12:00:48.896676Z",
     "iopub.status.idle": "2021-02-06T12:00:49.988747Z",
     "shell.execute_reply": "2021-02-06T12:00:49.988198Z"
    },
    "papermill": {
     "duration": 1.126335,
     "end_time": "2021-02-06T12:00:49.988891",
     "exception": false,
     "start_time": "2021-02-06T12:00:48.862556",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Loading the required libraries\n",
    "from lifetimes import BetaGeoFitter\n",
    "from lifetimes import GammaGammaFitter\n",
    "from lifetimes.plotting import plot_probability_alive_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:50.056479Z",
     "iopub.status.busy": "2021-02-06T12:00:50.052272Z",
     "iopub.status.idle": "2021-02-06T12:00:50.060975Z",
     "shell.execute_reply": "2021-02-06T12:00:50.061437Z"
    },
    "papermill": {
     "duration": 0.043956,
     "end_time": "2021-02-06T12:00:50.061614",
     "exception": false,
     "start_time": "2021-02-06T12:00:50.017658",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          Monetary  Recency\n",
      "Monetary   1.00000  0.09337\n",
      "Recency    0.09337  1.00000\n"
     ]
    }
   ],
   "source": [
    "# Checking BG/NBD model assumption and requirements\n",
    "print(cltv[['Monetary', 'Recency']].corr())  # Correlation between monetary ve recency variables\n",
    "cltv[\"Frequency\"] = cltv[\"Frequency\"].astype(int) # Type of frequency variable should be integer for BG-NBD model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:50.127395Z",
     "iopub.status.busy": "2021-02-06T12:00:50.126646Z",
     "iopub.status.idle": "2021-02-06T12:00:50.293680Z",
     "shell.execute_reply": "2021-02-06T12:00:50.293140Z"
    },
    "papermill": {
     "duration": 0.201644,
     "end_time": "2021-02-06T12:00:50.293813",
     "exception": false,
     "start_time": "2021-02-06T12:00:50.092169",
     "status": "completed"
    },
    "scrolled": false,
    "tags": []
   },
   "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>Recency</th>\n",
       "      <th>T</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Monetary</th>\n",
       "      <th>expctd_num_of_purch</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>12748.0</th>\n",
       "      <td>372</td>\n",
       "      <td>374</td>\n",
       "      <td>210</td>\n",
       "      <td>154.192429</td>\n",
       "      <td>164.006741</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14911.0</th>\n",
       "      <td>372</td>\n",
       "      <td>374</td>\n",
       "      <td>201</td>\n",
       "      <td>691.710100</td>\n",
       "      <td>157.048730</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17841.0</th>\n",
       "      <td>371</td>\n",
       "      <td>374</td>\n",
       "      <td>124</td>\n",
       "      <td>330.134355</td>\n",
       "      <td>97.454047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13089.0</th>\n",
       "      <td>366</td>\n",
       "      <td>370</td>\n",
       "      <td>97</td>\n",
       "      <td>606.362474</td>\n",
       "      <td>77.200939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14606.0</th>\n",
       "      <td>372</td>\n",
       "      <td>374</td>\n",
       "      <td>93</td>\n",
       "      <td>130.139032</td>\n",
       "      <td>73.528328</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Recency    T  Frequency    Monetary  expctd_num_of_purch\n",
       "Customer ID                                                          \n",
       "12748.0          372  374        210  154.192429           164.006741\n",
       "14911.0          372  374        201  691.710100           157.048730\n",
       "17841.0          371  374        124  330.134355            97.454047\n",
       "13089.0          366  370         97  606.362474            77.200939\n",
       "14606.0          372  374         93  130.139032            73.528328"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating BG-NBD Model\n",
    "bgf = BetaGeoFitter(penalizer_coef=0.001) # model object\n",
    "bgf.fit(cltv['Frequency'], cltv['Recency'], cltv['T']) # model fitting\n",
    "\n",
    "# Prediction of expected number of transaction for each customer for one year (365 days)\n",
    "cltv['expctd_num_of_purch'] = bgf.predict(365, cltv['Frequency'], cltv['Recency'], cltv['T']) \n",
    "cltv.sort_values(\"expctd_num_of_purch\",ascending=False).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.028871,
     "end_time": "2021-02-06T12:00:50.353035",
     "exception": false,
     "start_time": "2021-02-06T12:00:50.324164",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Look at the new columns that we estimate average spending for each customer in a year. By using this information, you can make many business decisions!\n",
    "\n",
    "This model can also estimate the retention rate and the probability of the customer is alive. Let's plot the probability matrix together and see what will happen!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:50.433160Z",
     "iopub.status.busy": "2021-02-06T12:00:50.418716Z",
     "iopub.status.idle": "2021-02-06T12:00:50.815597Z",
     "shell.execute_reply": "2021-02-06T12:00:50.815110Z"
    },
    "papermill": {
     "duration": 0.433726,
     "end_time": "2021-02-06T12:00:50.815738",
     "exception": false,
     "start_time": "2021-02-06T12:00:50.382012",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "# set figure size\n",
    "plt.subplots(figsize=(10, 5))\n",
    "plot_probability_alive_matrix(bgf)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.030147,
     "end_time": "2021-02-06T12:00:50.876371",
     "exception": false,
     "start_time": "2021-02-06T12:00:50.846224",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# 5. Train Gamma Gamma Model (Expected average spending)\n",
    "\n",
    "Now, we will train the gamma-gamma model to estimate average spending for future transactional events. Then, it will be used for calculating the customer lifetime value. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:50.945455Z",
     "iopub.status.busy": "2021-02-06T12:00:50.944791Z",
     "iopub.status.idle": "2021-02-06T12:00:51.013319Z",
     "shell.execute_reply": "2021-02-06T12:00:51.012705Z"
    },
    "papermill": {
     "duration": 0.106405,
     "end_time": "2021-02-06T12:00:51.013460",
     "exception": false,
     "start_time": "2021-02-06T12:00:50.907055",
     "status": "completed"
    },
    "scrolled": true,
    "tags": []
   },
   "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>Recency</th>\n",
       "      <th>T</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Monetary</th>\n",
       "      <th>expctd_num_of_purch</th>\n",
       "      <th>expct_avg_spend</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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12347.0</th>\n",
       "      <td>365</td>\n",
       "      <td>368</td>\n",
       "      <td>7</td>\n",
       "      <td>615.714286</td>\n",
       "      <td>7.081226</td>\n",
       "      <td>631.911974</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12348.0</th>\n",
       "      <td>282</td>\n",
       "      <td>359</td>\n",
       "      <td>4</td>\n",
       "      <td>442.695000</td>\n",
       "      <td>4.602317</td>\n",
       "      <td>463.745539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12352.0</th>\n",
       "      <td>260</td>\n",
       "      <td>297</td>\n",
       "      <td>8</td>\n",
       "      <td>219.542500</td>\n",
       "      <td>9.090603</td>\n",
       "      <td>224.886669</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12356.0</th>\n",
       "      <td>302</td>\n",
       "      <td>326</td>\n",
       "      <td>3</td>\n",
       "      <td>937.143333</td>\n",
       "      <td>4.298322</td>\n",
       "      <td>995.997679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12358.0</th>\n",
       "      <td>149</td>\n",
       "      <td>151</td>\n",
       "      <td>2</td>\n",
       "      <td>575.210000</td>\n",
       "      <td>5.953164</td>\n",
       "      <td>631.900951</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Recency    T  Frequency    Monetary  expctd_num_of_purch  \\\n",
       "Customer ID                                                             \n",
       "12347.0          365  368          7  615.714286             7.081226   \n",
       "12348.0          282  359          4  442.695000             4.602317   \n",
       "12352.0          260  297          8  219.542500             9.090603   \n",
       "12356.0          302  326          3  937.143333             4.298322   \n",
       "12358.0          149  151          2  575.210000             5.953164   \n",
       "\n",
       "             expct_avg_spend  \n",
       "Customer ID                   \n",
       "12347.0           631.911974  \n",
       "12348.0           463.745539  \n",
       "12352.0           224.886669  \n",
       "12356.0           995.997679  \n",
       "12358.0           631.900951  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating Gamma-Gamma Model\n",
    "ggf = GammaGammaFitter(penalizer_coef=0.01) # model object\n",
    "ggf.fit(cltv['Frequency'], cltv['Monetary']) # model fitting\n",
    "\n",
    "# Prediction of expected amount of average profit\n",
    "cltv[\"expct_avg_spend\"] = ggf.conditional_expected_average_profit(cltv['Frequency'], cltv['Monetary'])\n",
    "\n",
    "cltv.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.030653,
     "end_time": "2021-02-06T12:00:51.075574",
     "exception": false,
     "start_time": "2021-02-06T12:00:51.044921",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## 6. Final: Calculate CLTV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-02-06T12:00:51.143176Z",
     "iopub.status.busy": "2021-02-06T12:00:51.142574Z",
     "iopub.status.idle": "2021-02-06T12:00:51.389334Z",
     "shell.execute_reply": "2021-02-06T12:00:51.388830Z"
    },
    "papermill": {
     "duration": 0.282943,
     "end_time": "2021-02-06T12:00:51.389483",
     "exception": false,
     "start_time": "2021-02-06T12:00:51.106540",
     "status": "completed"
    },
    "tags": []
   },
   "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>Recency</th>\n",
       "      <th>T</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Monetary</th>\n",
       "      <th>expctd_num_of_purch</th>\n",
       "      <th>expct_avg_spend</th>\n",
       "      <th>cltv_one_year</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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14646.0</th>\n",
       "      <td>353</td>\n",
       "      <td>355</td>\n",
       "      <td>74</td>\n",
       "      <td>3596.804392</td>\n",
       "      <td>61.316101</td>\n",
       "      <td>3605.309159</td>\n",
       "      <td>204735.095545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18102.0</th>\n",
       "      <td>366</td>\n",
       "      <td>368</td>\n",
       "      <td>60</td>\n",
       "      <td>3859.739083</td>\n",
       "      <td>48.623095</td>\n",
       "      <td>3870.996702</td>\n",
       "      <td>174313.470352</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14096.0</th>\n",
       "      <td>97</td>\n",
       "      <td>102</td>\n",
       "      <td>17</td>\n",
       "      <td>3163.588235</td>\n",
       "      <td>35.234619</td>\n",
       "      <td>3196.435385</td>\n",
       "      <td>104394.666897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12415.0</th>\n",
       "      <td>313</td>\n",
       "      <td>338</td>\n",
       "      <td>21</td>\n",
       "      <td>5724.302619</td>\n",
       "      <td>19.005586</td>\n",
       "      <td>5772.177190</td>\n",
       "      <td>101605.295723</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14911.0</th>\n",
       "      <td>372</td>\n",
       "      <td>374</td>\n",
       "      <td>201</td>\n",
       "      <td>691.710100</td>\n",
       "      <td>157.048730</td>\n",
       "      <td>692.326355</td>\n",
       "      <td>100694.105622</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Recency    T  Frequency     Monetary  expctd_num_of_purch  \\\n",
       "Customer ID                                                              \n",
       "14646.0          353  355         74  3596.804392            61.316101   \n",
       "18102.0          366  368         60  3859.739083            48.623095   \n",
       "14096.0           97  102         17  3163.588235            35.234619   \n",
       "12415.0          313  338         21  5724.302619            19.005586   \n",
       "14911.0          372  374        201   691.710100           157.048730   \n",
       "\n",
       "             expct_avg_spend  cltv_one_year  \n",
       "Customer ID                                  \n",
       "14646.0          3605.309159  204735.095545  \n",
       "18102.0          3870.996702  174313.470352  \n",
       "14096.0          3196.435385  104394.666897  \n",
       "12415.0          5772.177190  101605.295723  \n",
       "14911.0           692.326355  100694.105622  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculating customer lifetime value by using BG-NBD and GammaGamma models: \n",
    "\n",
    "cltv[\"cltv_one_year\"] = ggf.customer_lifetime_value(bgf,\n",
    "                                   cltv['Frequency'],\n",
    "                                   cltv['Recency'],\n",
    "                                   cltv['T'],\n",
    "                                   cltv['Monetary'],\n",
    "                                   time=12,  # 12 month\n",
    "                                   freq=\"D\",  # frequency of T\n",
    "                                   discount_rate=0.01)\n",
    "\n",
    "cltv.sort_values(\"cltv_one_year\",ascending=False).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.031297,
     "end_time": "2021-02-06T12:00:51.452589",
     "exception": false,
     "start_time": "2021-02-06T12:00:51.421292",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "<p>Together, we get the most valuable five customers for a year! </p>\n",
    "<p>This model is capable of producing valuable information to be used for several purposes by various departments such as marketing, sales, and product development. By using it, we can simply create segments according to the CLTV. We can also extract the least valuable customers and send them to customer relationship management to increase customer engagement with the company."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.031215,
     "end_time": "2021-02-06T12:00:51.515156",
     "exception": false,
     "start_time": "2021-02-06T12:00:51.483941",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Thank you for making it to the very end of this notebook!\n",
    "\n",
    "Please feel free to leave a comment to improve it together. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "papermill": {
     "duration": 0.031173,
     "end_time": "2021-02-06T12:00:51.577792",
     "exception": false,
     "start_time": "2021-02-06T12:00:51.546619",
     "status": "completed"
    },
    "tags": []
   },
   "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"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 116.93758,
   "end_time": "2021-02-06T12:00:53.317611",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2021-02-06T11:58:56.380031",
   "version": "2.2.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}