bgnbd_gammagamma_modeling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Author**: [Alex FitzGerald](https://www.linkedin.com/in/alex-fitzgerald-0734076a/)\n",
    "\n",
    "![windjammer header](visuals/windjammer_logo.jpg)\n",
    "\n",
    "## Overview\n",
    "This project creates a user-friendly customer lifetime value (CLV) prediction engine able to take in transaction data and return important CLV predictions with a high degree of accuracy for a merchant's entire customer base and individual customers over a selected period of time in the future. The data processing and modeling techniques were validated on a publicly available e-commerce data set from Olist but the app is flexible enough to re-process and model unstructured transaction data from any non-contractual business with continuous purchasing behavior (ex. grocery purchases, e-commerce, movie rentals).\n",
    "\n",
    "## Business Problem\n",
    "How much are your customers worth? CLV is a north star metric for marketers and key to many business analyses; sales/marketing activity ROI, sales forecasts, company valuation, churn mitigation tactics, and customer persona analysis. \n",
    "\n",
    "Olist, the largest e-commerce department store in Brazil, wants to help the small business merchants who sell goods through their marketplace understand their CLV. Olist makes their money taking a commission on orders placed on their marketplace. They want to help their merchants sell more products and retain those merchants on their marketplace by providing top-notch business analytics.\n",
    "\n",
    "Olist has decided to create a user friendly app to help their merchants better understand the customers who buy their products through the Olist marketplace. Empowering their merchants with CLV analysis will help merchants improve their strategy and increase sales and commissions through Olist. The addition of a customer insight tool will also improve the retention of merchants on the Olist network. \n",
    "\n",
    "To accomplish this project, Olist has hired Windjammer Consulting, a data-science consulting firm.\n",
    "\n",
    "To test out the app for accuracy, Olist has provided a full transaction data set of purchases from all vendors and asked Windjammer Consulting to return CLV predictions with minimal error (RMSE).\n",
    "If the model passes muster, they want Windjammer Consulting to create an application that allows merchants to upload their own data and retrieve predictions.\n",
    "\n",
    "## Modeling\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Inital imports\n",
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import lifetimes\n",
    "from lifetimes.plotting import plot_period_transactions, plot_calibration_purchases_vs_holdout_purchases\n",
    "from lifetimes import BetaGeoFitter, GammaGammaFitter\n",
    "from datetime import timedelta\n",
    "from datetime import datetime\n",
    "from dateutil import parser\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "OnlineRetail_2yrs.csv             customer_data_base.zip\r\n",
      "brave_transactions.csv            lifetimes_object_df.csv\r\n",
      "business_context.jpg              lifetimes_object_df_uncleaned.csv\r\n"
     ]
    }
   ],
   "source": [
    "!ls data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load in our pre-prepared dataframes\n",
    "lifetimes_object_df = pd.read_csv(\"data/lifetimes_object_df.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preparation and evaluation functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load in our data wrangler\n",
    "from lifetimes.utils import summary_data_from_transaction_data\n",
    "#Import holdout \n",
    "from lifetimes.utils import calibration_and_holdout_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#class to return a calibration and holdout df\n",
    "class df_ch():\n",
    "    def __init__(self, eval_period=None,transaction_df=None,purchase_timestamp_col=None,customer_id_col=None,datetime_col=None,monetary_value_col=None):\n",
    "        #initialized attributes\n",
    "        self.eval_period = eval_period\n",
    "        self.transaction_df = transaction_df\n",
    "        self.purchase_timestamp_col=purchase_timestamp_col\n",
    "        self.customer_id_col=customer_id_col\n",
    "        self.datetime_col=datetime_col\n",
    "        self.monetary_value_col=monetary_value_col\n",
    "        #save off more attributes\n",
    "        self.min_obs_date = parser.parse(transaction_df[purchase_timestamp_col].min())\n",
    "        self.max_obs_date = parser.parse(transaction_df[purchase_timestamp_col].max())\n",
    "        self.max_calib_date = self.max_obs_date - timedelta(days=eval_period)  \n",
    "        self.calib_range_days = (self.max_calib_date - self.min_obs_date).days\n",
    "    def df_ch_getdf(self):\n",
    "        df = calibration_and_holdout_data(\n",
    "        transactions = self.transaction_df, \n",
    "        customer_id_col=self.customer_id_col,\n",
    "        datetime_col=self.datetime_col,\n",
    "        monetary_value_col=self.monetary_value_col,\n",
    "        calibration_period_end = self.max_calib_date, \n",
    "        observation_period_end = self.max_obs_date, \n",
    "        freq = \"D\")\n",
    "        return df\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#function to return a df of real and predicted transacitons in eval period\n",
    "def bgf_real_v_pred_df(ch,bgf):\n",
    "    rfm_cal_holdout = pd.DataFrame()\n",
    "    ch_df = ch.df_ch_getdf()\n",
    "    rfm_cal_holdout[\"n_transactions_cal_real\"]  = ch_df[\"frequency_cal\"] + 1 #Total calibration days with purchases = calibration frequency + 1\n",
    "    rfm_cal_holdout[\"n_transactions_holdout_real\"]  = ch_df[\"frequency_holdout\"] #Total validation days with purchases = validation frequency\n",
    "    # the predicted number of transactions\n",
    "    rfm_cal_holdout[\"n_transactions_holdout_pred\"] = bgf.predict(t=ch.eval_period, \n",
    "                                                    frequency=ch_df['frequency_cal'], \n",
    "                                                    recency=ch_df['recency_cal'], \n",
    "                                                    T=ch_df['T_cal'])\n",
    "    return rfm_cal_holdout[[\"n_transactions_cal_real\",\"n_transactions_holdout_real\", \"n_transactions_holdout_pred\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#function to capture RMSE for a BGF model\n",
    "def bgf_rmse(ch,bgf):\n",
    "    df_ch = ch.df_ch_getdf()\n",
    "    df_ch[\"n_transactions_holdout_real\"] = df_ch[\"frequency_holdout\"]\n",
    "    y_true = df_ch[\"n_transactions_holdout_real\"]\n",
    "    y_pred = bgf.predict(t=ch.eval_period, frequency=df_ch['frequency_cal'],\n",
    "                         recency=df_ch['recency_cal'],\n",
    "                         T=df_ch['T_cal'])\n",
    "\n",
    "    return mean_squared_error(y_true,y_pred,squared=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#function to return predicted # transactions for given customer in evaluation period\n",
    "def samp_cust_pred_trans(df_ch,sample_customer_id,eval_period):\n",
    "    sample_customer = df_ch.loc[sample_customer_id]\n",
    "    n_transactions_pred = bgf.predict(t=eval_period,\n",
    "                                  frequency=sample_customer['frequency_cal'], \n",
    "                                  recency=sample_customer['recency_cal'], \n",
    "                                  T=sample_customer['T_cal'])\n",
    "    return(n_transactions_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Dummy model function that predicts the same purchase rate in the observation period to continue through the eval period\n",
    "def bgf_dummy_model(ch):\n",
    "    df_ch = ch.df_ch_getdf()\n",
    "    dummy_ch_df = pd.DataFrame()\n",
    "    dummy_ch_df[\"purchases_per_period_cal\"] = df_ch[\"frequency_cal\"] + 1 #Total calibration days with purchases = calibration frequency + 1\n",
    "    dummy_ch_df[\"purchases_per_period_cal\"] = dummy_ch_df[\"purchases_per_period_cal\"]/ch.calib_range_days #Purchase rate\n",
    "    dummy_ch_df[\"dummy_pred_purchases_holdout\"] = dummy_ch_df[\"purchases_per_period_cal\"] * ch.eval_period #Predicting purchase rate to continue\n",
    "    dummy_ch_df[\"dummy_pred_purchases_holdout\"] = np.round(dummy_ch_df[\"dummy_pred_purchases_holdout\"]) #Round to get an actual number of predictions\n",
    "    dummy_ch_df[\"actual_purchases_holdout\"] = df_ch[\"frequency_holdout\"] #Total validation days with purchases = validation frequency\n",
    "    return mean_squared_error(y_true=dummy_ch_df[\"dummy_pred_purchases_holdout\"],y_pred=dummy_ch_df[\"actual_purchases_holdout\"],squared=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#function to return a df of real and predicted transacitons in eval period for dummy model\n",
    "def bgf_dummy_real_v_pred_df(ch):\n",
    "    df_ch = ch.df_ch_getdf()\n",
    "    dummy_ch_df = pd.DataFrame()\n",
    "    dummy_ch_df[\"purchases_per_period_cal\"] = df_ch[\"frequency_cal\"] + 1 #Total calibration days with purchases = calibration frequency + 1\n",
    "    dummy_ch_df[\"purchases_per_period_cal\"] = dummy_ch_df[\"purchases_per_period_cal\"]/ch.calib_range_days #Purchase rate\n",
    "    dummy_ch_df[\"dummy_pred_purchases_holdout\"] = dummy_ch_df[\"purchases_per_period_cal\"] * ch.eval_period #Predicting purchase rate to continue\n",
    "    dummy_ch_df[\"actual_purchases_holdout\"] = df_ch[\"frequency_holdout\"] #Total validation days with purchases = validation frequency\n",
    "    return dummy_ch_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### RFM data exploration "
   ]
  },
  {
   "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>frequency</th>\n",
       "      <th>recency</th>\n",
       "      <th>T</th>\n",
       "      <th>monetary_value</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0000366f3b9a7992bf8c76cfdf3221e2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>160.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000b849f77a49e4a4ce2b2a4ca5be3f</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>163.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000f46a3911fa3c0805444483337064</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>586.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0000f6ccb0745a6a4b88665a16c9f078</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>370.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0004aac84e0df4da2b147fca70cf8255</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>337.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fffcf5a5ff07b0908bd4e2dbc735a684</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>496.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fffea47cd6d3cc0a88bd621562a9d061</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>311.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ffff371b4d645b6ecea244b27531430a</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>617.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ffff5962728ec6157033ef9805bacc48</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>168.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ffffd2657e2aad2907e67c3e9daecbeb</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>533.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>94989 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  frequency  recency      T  monetary_value\n",
       "customer_unique_id                                                         \n",
       "0000366f3b9a7992bf8c76cfdf3221e2        0.0      0.0  160.0             0.0\n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f        0.0      0.0  163.0             0.0\n",
       "0000f46a3911fa3c0805444483337064        0.0      0.0  586.0             0.0\n",
       "0000f6ccb0745a6a4b88665a16c9f078        0.0      0.0  370.0             0.0\n",
       "0004aac84e0df4da2b147fca70cf8255        0.0      0.0  337.0             0.0\n",
       "...                                     ...      ...    ...             ...\n",
       "fffcf5a5ff07b0908bd4e2dbc735a684        0.0      0.0  496.0             0.0\n",
       "fffea47cd6d3cc0a88bd621562a9d061        0.0      0.0  311.0             0.0\n",
       "ffff371b4d645b6ecea244b27531430a        0.0      0.0  617.0             0.0\n",
       "ffff5962728ec6157033ef9805bacc48        0.0      0.0  168.0             0.0\n",
       "ffffd2657e2aad2907e67c3e9daecbeb        0.0      0.0  533.0             0.0\n",
       "\n",
       "[94989 rows x 4 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#use Lifetimes utilitity function to transform our data into an RFM (recency, requency, monetary value) object\n",
    "summary = summary_data_from_transaction_data(transactions=lifetimes_object_df,customer_id_col='customer_unique_id',\n",
    "                                             datetime_col='order_purchase_timestamp',monetary_value_col='payment_value',observation_period_end='2018-10-17')\n",
    "summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>frequency</th>\n",
       "      <th>recency</th>\n",
       "      <th>T</th>\n",
       "      <th>monetary_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>frequency</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.702598</td>\n",
       "      <td>0.069469</td>\n",
       "      <td>0.614334</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>recency</th>\n",
       "      <td>0.702598</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.101774</td>\n",
       "      <td>0.472158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>T</th>\n",
       "      <td>0.069469</td>\n",
       "      <td>0.101774</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.044082</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>monetary_value</th>\n",
       "      <td>0.614334</td>\n",
       "      <td>0.472158</td>\n",
       "      <td>0.044082</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                frequency   recency         T  monetary_value\n",
       "frequency        1.000000  0.702598  0.069469        0.614334\n",
       "recency          0.702598  1.000000  0.101774        0.472158\n",
       "T                0.069469  0.101774  1.000000        0.044082\n",
       "monetary_value   0.614334  0.472158  0.044082        1.000000"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#We have some significant co-linearity but this isn't a concern. We actually expect co-lineary which is why we are\n",
    "#using a hierarchical modeling approach because we expect groups with common rfm scores to move togiether. \n",
    "#Importantly, we do not have co-linearity between T and monetary_value\n",
    "summary.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0     0.978261\n",
       "1.0     0.020097\n",
       "2.0     0.001305\n",
       "3.0     0.000189\n",
       "4.0     0.000063\n",
       "5.0     0.000053\n",
       "6.0     0.000021\n",
       "15.0    0.000011\n",
       "Name: frequency, dtype: float64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#We have less than 3% of customers who are repeat purchasers\n",
    "summary.frequency.value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2065"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Looking just at customers who have placed more than 1 order\n",
    "len(summary)- summary.frequency.value_counts()[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "customers without repeat purchases: 92924\n"
     ]
    }
   ],
   "source": [
    "#These customers cannot be used for analysis in your GG model bc they don't have repeat purchases\n",
    "print(f'customers without repeat purchases: {summary.frequency.value_counts()[0]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#create a list of repeat purchasers\n",
    "repeat_purchaser_ids = list(summary[summary[\"frequency\"] >= 1].index)\n",
    "#subset our transaction data to just repeat purchasers\n",
    "repeat_lifetimes_object_df = lifetimes_object_df[lifetimes_object_df[\"customer_unique_id\"].isin(repeat_purchaser_ids) == True]\n",
    "#repeat_lifetimes_object_df[\"order_purchase_timestamp\"] = pd.to_datetime(repeat_lifetimes_object_df[\"order_purchase_timestamp\"])\n",
    "#Checking we removed the correct number of user id's\n",
    "(len(lifetimes_object_df.customer_unique_id.unique()) - len(repeat_lifetimes_object_df.customer_unique_id.unique())) == summary.frequency.value_counts()[0]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "repeat_purchaser_summary = summary_data_from_transaction_data(transactions=repeat_lifetimes_object_df,customer_id_col='customer_unique_id',\n",
    "                                             datetime_col='order_purchase_timestamp',monetary_value_col='payment_value')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "113.9720823244552"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Calculate the avg time between purchases\n",
    "inter_purchase_period = []\n",
    "for customer in repeat_purchaser_summary.index:\n",
    "    inter_purchase_period.append((repeat_purchaser_summary.loc[customer].recency/repeat_purchaser_summary.loc[customer].frequency))\n",
    "avg_inter_purchase_period = sum(inter_purchase_period)/len(inter_purchase_period)\n",
    "avg_inter_purchase_period"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "700"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#From the documentation, for optimal fit, the training date range should cover at least 3 \n",
    "#but hopefully 5-10 inter-purchase periods\n",
    "#The holdout period should be at least half the training period\n",
    "total_date_range = (pd.to_datetime(repeat_lifetimes_object_df[\"order_purchase_timestamp\"]).max() - pd.to_datetime(repeat_lifetimes_object_df[\"order_purchase_timestamp\"]).min()).days\n",
    "total_date_range"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "holdout period given 4 inter period multiple: 244.1116707021792\n"
     ]
    }
   ],
   "source": [
    "# An inter_period_multiple of 4 works well and gives us a holdout period of about 240 days\n",
    "inter_period_multiple = 4\n",
    "print(f'holdout period given {inter_period_multiple} inter period multiple: {total_date_range-(avg_inter_purchase_period*4)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>monetary_value</th>\n",
       "      <th>frequency</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>monetary_value</th>\n",
       "      <td>1.00000</td>\n",
       "      <td>-0.00049</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>frequency</th>\n",
       "      <td>-0.00049</td>\n",
       "      <td>1.00000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                monetary_value  frequency\n",
       "monetary_value         1.00000   -0.00049\n",
       "frequency             -0.00049    1.00000"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#The Gamma-Gamma submodel, assumes no relationship between the monetary value and the purchase frequency. \n",
    "#The correlation between the two vectors is close to 0 so we can use this model.\n",
    "repeat_purchaser_summary[['monetary_value', 'frequency']].corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load in our ch object with all data\n",
    "ch_1 = df_ch(eval_period=240,transaction_df=lifetimes_object_df,purchase_timestamp_col='order_purchase_timestamp',\n",
    "                   customer_id_col='customer_unique_id',datetime_col='order_purchase_timestamp',\n",
    "                   monetary_value_col='payment_value')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load in our ch object with repeat purchaser data\n",
    "ch_2 = df_ch(eval_period=240,transaction_df=repeat_lifetimes_object_df,purchase_timestamp_col='order_purchase_timestamp',\n",
    "                   customer_id_col='customer_unique_id',datetime_col='order_purchase_timestamp',\n",
    "                   monetary_value_col='payment_value')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "#create an array to evaluate total purchases per customer in calib period\n",
    "lifetimes_object_df_ts = lifetimes_object_df\n",
    "lifetimes_object_df_ts[\"order_purchase_timestamp\"] = pd.to_datetime(lifetimes_object_df_ts[\"order_purchase_timestamp\"])\n",
    "purchases_in_cal = lifetimes_object_df_ts[lifetimes_object_df_ts[\"order_purchase_timestamp\"]<=ch_1.max_calib_date]\n",
    "purchases_in_cal = purchases_in_cal[purchases_in_cal[\"order_purchase_timestamp\"]<=ch_1.max_calib_date].groupby(\"customer_unique_id\").nunique()[\"order_purchase_timestamp\"]\n",
    "purchases_in_cal.sort_values()\n",
    "purchases_in_cal = pd.DataFrame(purchases_in_cal)\n",
    "purchaser_id_in_cal = list(purchases_in_cal.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "#create an array to evaluate total purchases per customer in validation period\n",
    "purchases_in_val = lifetimes_object_df_ts[lifetimes_object_df_ts[\"order_purchase_timestamp\"]>ch_1.max_calib_date].groupby(\"customer_unique_id\").nunique()[\"order_purchase_timestamp\"]\n",
    "purchases_in_val.sort_values()\n",
    "purchases_in_val = pd.DataFrame(purchases_in_val)\n",
    "#subset to just those id's that appeared in our calib period\n",
    "purchases_in_val = purchases_in_val[purchases_in_val.index.isin(purchaser_id_in_cal)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Get ch df from repeat purchaser ch object\n",
    "df_ch_2 = ch_2.df_ch_getdf()\n",
    "#Iniatialze df_ch for all purchasers\n",
    "df_ch_1 = ch_1.df_ch_getdf()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Confirming that a 'frequency' of 0.00 in calibration period means total purchases in that period = 1\n",
    "#Total calibration days with purchases = calibration frequency + 1\n",
    "purch = []\n",
    "for purchasers in df_ch_1[\"frequency_cal\"]:    \n",
    "    purch.append(purchasers+1)\n",
    "sum(purch) == purchases_in_cal[\"order_purchase_timestamp\"].sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Attempting on holdout group\n",
    "#False bc these are repeat customers so we don't need to add +1\n",
    "purch = []\n",
    "for purchasers in df_ch_1[\"frequency_holdout\"]:    \n",
    "    purch.append(purchasers+1)\n",
    "sum(purch) == purchases_in_val[\"order_purchase_timestamp\"].sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Total validation days with purchases = validation frequency\n",
    "sum(df_ch_1[\"frequency_holdout\"]) == purchases_in_val[\"order_purchase_timestamp\"].sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Dummy model\n",
    "Model features\n",
    "- Predict purchases in holdout period based on the observed purchase rate in the calibration period\n",
    "\n",
    "Model results\n",
    "- Decent RMSE for all customers, the class imbalence (most customers never re-ordder) helps out it's niave assumptions\n",
    "- Poor RMSE for repeat customer subset\n",
    "- The dummy model expects purchase rate to continue so you see it do very poorly predicting actual purchase rate for customers who had just one purchase in the calibration period."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " bgf dummy model RMSE w/ all data: 0.18670952242425554\n",
      " bgf dummy model RMSE w/ repeat cust only: 0.8206158691033689\n"
     ]
    }
   ],
   "source": [
    "print(f' bgf dummy model RMSE w/ all data: {bgf_dummy_model(ch_1)}')\n",
    "print(f' bgf dummy model RMSE w/ repeat cust only: {bgf_dummy_model(ch_2)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Return a dataframe with predicted and actual values using dummy model on repeat customers\n",
    "dummy_ch_df = pd.DataFrame()\n",
    "dummy_ch_df[\"frequency_cal\"] = df_ch_1[\"frequency_cal\"]\n",
    "#Add 1 to frequency to get total purchasers in calibration period\n",
    "dummy_ch_df[\"purchases_per_period_cal\"] = df_ch_1[\"frequency_cal\"] + 1\n",
    "#divide by purchases by n calibration days to get purchase rate\n",
    "dummy_ch_df[\"purchases_per_period_cal\"] = dummy_ch_df[\"purchases_per_period_cal\"]/ch_1.calib_range_days\n",
    "#multiply purchase rate by n evaluation days to get purchase per day\n",
    "dummy_ch_df[\"dummy_pred_purchases_holdout\"] = dummy_ch_df[\"purchases_per_period_cal\"] * ch_1.eval_period\n",
    "#Round the prediction to get an integer, purchases either happen or they don't\n",
    "dummy_ch_df[\"dummy_pred_purchases_holdout\"] = np.round(dummy_ch_df[\"dummy_pred_purchases_holdout\"])\n",
    "#Get actual purchases in validation period\n",
    "dummy_ch_df[\"actual_purchases_holdout\"] = df_ch_1[\"frequency_holdout\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>purchases_per_period_cal</th>\n",
       "      <th>dummy_pred_purchases_holdout</th>\n",
       "      <th>actual_purchases_holdout</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>frequency_cal</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.0</th>\n",
       "      <td>43702</td>\n",
       "      <td>43702</td>\n",
       "      <td>43702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>711</td>\n",
       "      <td>711</td>\n",
       "      <td>711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2.0</th>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3.0</th>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4.0</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8.0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               purchases_per_period_cal  dummy_pred_purchases_holdout  \\\n",
       "frequency_cal                                                           \n",
       "0.0                               43702                         43702   \n",
       "1.0                                 711                           711   \n",
       "2.0                                  38                            38   \n",
       "3.0                                   9                             9   \n",
       "4.0                                   2                             2   \n",
       "8.0                                   1                             1   \n",
       "\n",
       "               actual_purchases_holdout  \n",
       "frequency_cal                            \n",
       "0.0                               43702  \n",
       "1.0                                 711  \n",
       "2.0                                  38  \n",
       "3.0                                   9  \n",
       "4.0                                   2  \n",
       "8.0                                   1  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#There are so few examples of customers with more than 4 purchases in the calibaration period that we should limit our\n",
    "#evaluation to 4 or fewer\n",
    "dummy_ch_df.groupby(\"frequency_cal\").count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot a visual to demonstrate dummy model performance\n",
    "fig, ax = plt.subplots(figsize=(10,5))\n",
    "ax.plot(dummy_ch_df.groupby(\"frequency_cal\").mean()[\"actual_purchases_holdout\"],label=('frequency_holdout'))\n",
    "ax.plot(dummy_ch_df.groupby(\"frequency_cal\").mean()[\"dummy_pred_purchases_holdout\"],label=('dummy_model_pred'))\n",
    "\n",
    "ax.set_xticks([0,1,2,3,4])\n",
    "ax.set_xlim(0,4)\n",
    "ax.set_yticks([0,1,2,3,4])\n",
    "ax.set_ylim(0,4)\n",
    "\n",
    "plt.xlabel(\"Purchases in calibration period\", fontsize=18)\n",
    "plt.ylabel(\"Purchases in holdout period\", fontsize=18)\n",
    "plt.xlabel(\"Purchases in calibration period\", fontsize=18)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.title(\"Dummy Model Performance\",fontsize=18)\n",
    "plt.legend([\"Actual frequency\",\"Model frequency\"],fontsize=18)\n",
    "\n",
    "#ax.legend()\n",
    "#Save figure as png\n",
    "plt.savefig(\"visuals/dummy_model_performance_repeat_purchasers.png\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  1.0 BG/NBD model\n",
    "Model features\n",
    "- All customers including those with just 1 purchase\n",
    "\n",
    "Model results\n",
    "- Very strong RMSE, this is our best model\n",
    "- We significantly improved upon our dummy model\n",
    "- Our model doesn't suffer from the niave assumption that purchase rate will just continue\n",
    "- Our predictions start to fail at higher frequencies but there are very few instances of these types of customers, a more balenced data set would aleviate this issue\n",
    "\n",
    "Next steps\n",
    "- Attempt on subsetted data of customers with more than 1 purchase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Iniatialize bgf model\n",
    "bgf_1 = BetaGeoFitter(penalizer_coef=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.065817\n",
      "         Iterations: 60\n",
      "         Function evaluations: 63\n",
      "         Gradient evaluations: 63\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<lifetimes.BetaGeoFitter: fitted with 44463 subjects, a: 1.85, alpha: 55.22, b: 0.32, r: 0.02>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Fit model to df_ch\n",
    "bgf_1.fit(\n",
    "        frequency = df_ch_1[\"frequency_cal\"], \n",
    "        recency = df_ch_1[\"recency_cal\"], \n",
    "        T = df_ch_1[\"T_cal\"],   \n",
    "        weights = None,  \n",
    "        verbose = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model is accurate to within 0.142 purchases over 240 days\n"
     ]
    }
   ],
   "source": [
    "#Return rmse for first model\n",
    "bgf_1_rmse = bgf_rmse(ch_1,bgf_1)\n",
    "print(f'Model is accurate to within {round(bgf_1_rmse,ndigits=3)} purchases over {int(ch_1.eval_period)} days')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>recency_cal</th>\n",
       "      <th>T_cal</th>\n",
       "      <th>monetary_value_cal</th>\n",
       "      <th>frequency_holdout</th>\n",
       "      <th>monetary_value_holdout</th>\n",
       "      <th>duration_holdout</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>frequency_cal</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.0</th>\n",
       "      <td>43702</td>\n",
       "      <td>43702</td>\n",
       "      <td>43702</td>\n",
       "      <td>43702</td>\n",
       "      <td>43702</td>\n",
       "      <td>43702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>711</td>\n",
       "      <td>711</td>\n",
       "      <td>711</td>\n",
       "      <td>711</td>\n",
       "      <td>711</td>\n",
       "      <td>711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2.0</th>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3.0</th>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4.0</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8.0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               recency_cal  T_cal  monetary_value_cal  frequency_holdout  \\\n",
       "frequency_cal                                                              \n",
       "0.0                  43702  43702               43702              43702   \n",
       "1.0                    711    711                 711                711   \n",
       "2.0                     38     38                  38                 38   \n",
       "3.0                      9      9                   9                  9   \n",
       "4.0                      2      2                   2                  2   \n",
       "8.0                      1      1                   1                  1   \n",
       "\n",
       "               monetary_value_holdout  duration_holdout  \n",
       "frequency_cal                                            \n",
       "0.0                             43702             43702  \n",
       "1.0                               711               711  \n",
       "2.0                                38                38  \n",
       "3.0                                 9                 9  \n",
       "4.0                                 2                 2  \n",
       "8.0                                 1                 1  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#There are so few examples of customers with more than 4 purchases in the calibaration period that we should limit our\n",
    "#evaluation to 4 or fewer\n",
    "df_ch_1.groupby(\"frequency_cal\").count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Applications/anaconda3/envs/lifetime-env/lib/python3.10/site-packages/lifetimes/generate_data.py:54: RuntimeWarning: divide by zero encountered in double_scalars\n",
      "  next_purchase_in = random.exponential(scale=1.0 / l)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Our model accurately esimates n purchases with in the calibation period\n",
    "plot_period_transactions(bgf_1);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# testing: predicted vs actual purchases in holdout period\n",
    "fig = plot_calibration_purchases_vs_holdout_purchases(bgf_1, df_ch_1,figsize=(10, 5))\n",
    "\n",
    "plt.xlabel(\"Purchases in calibration period\", fontsize=18)\n",
    "plt.ylabel(\"Purchases in holdout period\", fontsize=18)\n",
    "plt.xlabel(\"Purchases in calibration period\", fontsize=18)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.title(\"BG/NBD Model Performance\",fontsize=18)\n",
    "plt.legend([\"Actual frequency\",\"Model frequency\"],fontsize=18)\n",
    "\n",
    "fig.set_xticks([0,1,2,3,4])\n",
    "fig.set_xlim(0,4)\n",
    "fig.set_yticks([0,1,2,3,4])\n",
    "fig.set_ylim(0,4)\n",
    "#Save figure as png\n",
    "plt.savefig(\"visuals/final_bgf_model_performance.png\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  2.0 BG/NBD model\n",
    "Model features\n",
    "- Subsetted data of customers with more than 1 purchase\n",
    "\n",
    "Model results\n",
    "- Still pretty strong RMSE but not as good as the full data set\n",
    "- Our model is less accurate on this subset of repeat customers bc we can't inflate accuracy with the signficant n of correct 0 purchase perdictions that \n",
    "- That said, this model did much better than our dummy model did on repeat customers\n",
    "\n",
    "Next steps\n",
    "- Attempt regularization on the full data-set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Iniatialze df_ch \n",
    "df_ch_2 = ch_2.df_ch_getdf()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Iniatialize bgf model\n",
    "bgf_2 = BetaGeoFitter(penalizer_coef=0.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.095057\n",
      "         Iterations: 29\n",
      "         Function evaluations: 31\n",
      "         Gradient evaluations: 31\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<lifetimes.BetaGeoFitter: fitted with 1384 subjects, a: 1.82, alpha: 135.02, b: 0.25, r: 1.13>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Fit model to df_ch\n",
    "bgf_2.fit(\n",
    "        frequency = df_ch_2[\"frequency_cal\"],\n",
    "        recency = df_ch_2[\"recency_cal\"], \n",
    "        T = df_ch_2[\"T_cal\"],   \n",
    "        weights = None,  \n",
    "        verbose = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model is accurate to within 0.505 purchases over 240 days\n"
     ]
    }
   ],
   "source": [
    "bgf_2_rmse = bgf_rmse(ch_2,bgf_2)\n",
    "print(f'Model is accurate to within {round(bgf_2_rmse,ndigits=3)} purchases over {int(ch_2.eval_period)} days')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_period_transactions(bgf_2);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# testing: predicted vs actual purchases in holdout period \n",
    "fig = plot_calibration_purchases_vs_holdout_purchases(bgf_2, df_ch_2,figsize=(10, 5))\n",
    "\n",
    "plt.xlabel(\"Purchases in calibration period\", fontsize=18)\n",
    "plt.ylabel(\"Purchases in holdout period\", fontsize=18)\n",
    "plt.xlabel(\"Purchases in calibration period\", fontsize=18)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.title(\"BG/NBD Model Performance\",fontsize=18)\n",
    "plt.legend([\"Actual frequency\",\"Model frequency\"],fontsize=18)\n",
    "\n",
    "fig.set_xticks([0,1,2,3,4])\n",
    "fig.set_xlim(0,4)\n",
    "fig.set_yticks([0,1,2,3,4])\n",
    "fig.set_ylim(0,4);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  3.0 BG/NBD model\n",
    "Model features\n",
    "- All customers including those with just 1 purchase\n",
    "- Regularization in the form of l2 penalty\n",
    "\n",
    "Model results\n",
    "- Zero regularization produced the best result\n",
    "\n",
    "Next steps\n",
    "- Bring in a Gama Gama model to predict LTV using our best model `bgf_1` that has 0 regularization\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bgf_3_0 0.1423220657607575\n",
      "bgf_3_001 0.1469498734889828\n",
      "bgf_3_05 0.1639692144239852\n",
      "bgf_3_01 0.15561347485601168\n",
      "bgf_3_015 0.15753830802505678\n"
     ]
    }
   ],
   "source": [
    "#Documentation says that penalties in the order of 0.001 to 0.1 are effective\n",
    "penalizer_coef_grid = {\"bgf_3_0\":0,\"bgf_3_001\":0.001,\"bgf_3_05\":0.05,\"bgf_3_01\":.01,\"bgf_3_015\":.015}\n",
    "for penalizer in penalizer_coef_grid:\n",
    "    model = BetaGeoFitter(penalizer_coef=penalizer_coef_grid[penalizer])\n",
    "    model.fit(\n",
    "        frequency = df_ch_1[\"frequency_cal\"], \n",
    "        recency = df_ch_1[\"recency_cal\"], \n",
    "        T = df_ch_1[\"T_cal\"],   \n",
    "        weights = None,  \n",
    "        verbose = False)\n",
    "    print(f'{penalizer} {bgf_rmse(ch_1,model)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  1.0 GG model\n",
    "Model features\n",
    "- Customers with monetary values in training and validation periods\n",
    "- Minimal regularization\n",
    "\n",
    "Model results\n",
    "- Very good RMSE of around 6 which is only 0.05 standard deviations of our target variable\n",
    "- In comparision the dummy model was off by nearly 200 and 1.38 standard deviations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>frequency_cal</th>\n",
       "      <th>recency_cal</th>\n",
       "      <th>T_cal</th>\n",
       "      <th>monetary_value_cal</th>\n",
       "      <th>frequency_holdout</th>\n",
       "      <th>monetary_value_holdout</th>\n",
       "      <th>duration_holdout</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer_unique_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>041caba6a63ace5818387698ea85cdb2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>247.0</td>\n",
       "      <td>297.0</td>\n",
       "      <td>51.73</td>\n",
       "      <td>1.0</td>\n",
       "      <td>71.78</td>\n",
       "      <td>240.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>083ca1aa470c280236380973a48f77c6</th>\n",
       "      <td>2.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>162.0</td>\n",
       "      <td>86.96</td>\n",
       "      <td>1.0</td>\n",
       "      <td>105.29</td>\n",
       "      <td>240.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0b18593e033697810a62a987e77f4b8c</th>\n",
       "      <td>1.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>218.0</td>\n",
       "      <td>73.34</td>\n",
       "      <td>1.0</td>\n",
       "      <td>63.13</td>\n",
       "      <td>240.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1b6c7548a2a1f9037c1fd3ddfed95f33</th>\n",
       "      <td>4.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>54.0</td>\n",
       "      <td>138.31</td>\n",
       "      <td>2.0</td>\n",
       "      <td>176.81</td>\n",
       "      <td>240.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1da09dd64e235e7c2f29a4faff33535c</th>\n",
       "      <td>1.0</td>\n",
       "      <td>71.0</td>\n",
       "      <td>241.0</td>\n",
       "      <td>859.36</td>\n",
       "      <td>1.0</td>\n",
       "      <td>830.64</td>\n",
       "      <td>240.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  frequency_cal  recency_cal  T_cal  \\\n",
       "customer_unique_id                                                    \n",
       "041caba6a63ace5818387698ea85cdb2            1.0        247.0  297.0   \n",
       "083ca1aa470c280236380973a48f77c6            2.0         19.0  162.0   \n",
       "0b18593e033697810a62a987e77f4b8c            1.0         12.0  218.0   \n",
       "1b6c7548a2a1f9037c1fd3ddfed95f33            4.0         37.0   54.0   \n",
       "1da09dd64e235e7c2f29a4faff33535c            1.0         71.0  241.0   \n",
       "\n",
       "                                  monetary_value_cal  frequency_holdout  \\\n",
       "customer_unique_id                                                        \n",
       "041caba6a63ace5818387698ea85cdb2               51.73                1.0   \n",
       "083ca1aa470c280236380973a48f77c6               86.96                1.0   \n",
       "0b18593e033697810a62a987e77f4b8c               73.34                1.0   \n",
       "1b6c7548a2a1f9037c1fd3ddfed95f33              138.31                2.0   \n",
       "1da09dd64e235e7c2f29a4faff33535c              859.36                1.0   \n",
       "\n",
       "                                  monetary_value_holdout  duration_holdout  \n",
       "customer_unique_id                                                          \n",
       "041caba6a63ace5818387698ea85cdb2                   71.78             240.0  \n",
       "083ca1aa470c280236380973a48f77c6                  105.29             240.0  \n",
       "0b18593e033697810a62a987e77f4b8c                   63.13             240.0  \n",
       "1b6c7548a2a1f9037c1fd3ddfed95f33                  176.81             240.0  \n",
       "1da09dd64e235e7c2f29a4faff33535c                  830.64             240.0  "
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Subset our data-set to ensure there are monetary values in both calibration and holdout periods\n",
    "df_ch_2_subset = df_ch_2[df_ch_2[\"monetary_value_cal\"]>0]\n",
    "df_ch_2_subset = df_ch_2_subset[df_ch_2_subset[\"monetary_value_holdout\"]>0]\n",
    "df_ch_2_subset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<lifetimes.GammaGammaFitter: fitted with 51 subjects, p: 10.43, q: 1.27, v: 9.91>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Fitting our model using a minimal penalizer\n",
    "gg_1 = GammaGammaFitter(penalizer_coef = 0.001)\n",
    "gg_1.fit(frequency=df_ch_2_subset[\"frequency_cal\"],monetary_value=df_ch_2_subset[\"monetary_value_cal\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count     51.000000\n",
       "mean     138.458452\n",
       "std      136.872379\n",
       "min       43.713787\n",
       "25%       68.512605\n",
       "50%      100.035614\n",
       "75%      143.295964\n",
       "max      819.621051\n",
       "dtype: float64"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#returning and saving predicted average value for transactions\n",
    "predicted_monetary_value = gg_1.conditional_expected_average_profit(frequency=df_ch_2_subset[\"frequency_holdout\"],monetary_value=df_ch_2_subset[\"monetary_value_holdout\"])\n",
    "df_ch_2_subset['predicted_monetary_value'] = predicted_monetary_value\n",
    "predicted_monetary_value.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count     51.000000\n",
       "mean     132.894600\n",
       "std      140.075264\n",
       "min       34.920000\n",
       "25%       60.652143\n",
       "50%       92.680000\n",
       "75%      137.045000\n",
       "max      830.640000\n",
       "Name: monetary_value_holdout, dtype: float64"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Saving actual value std for evaluation purposes\n",
    "monetary_value_holdout_std = df_ch_2_subset[\"monetary_value_holdout\"].describe()[\"std\"]\n",
    "#our predicted values appear similar to the actual values!\n",
    "df_ch_2_subset[\"monetary_value_holdout\"].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Return mean squared error for predicted and a dummy model that assumes purchase value to continue from calibration to holdout exactly the same.\n",
    "gg_1_rmse = mean_squared_error(y_true=df_ch_2_subset[\"monetary_value_holdout\"],y_pred=predicted_monetary_value,squared=False)\n",
    "dummy_gg_rmse = mean_squared_error(y_true=df_ch_2_subset[\"monetary_value_holdout\"],y_pred=df_ch_2_subset[\"monetary_value_cal\"],squared=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " dummy gg model RMSE w/ repeat cust only: 192.78, which is equal to 1.38 standard deviations\n",
      " gg model RMSE w/ repeat cust only: 6.63, which is equal to 0.05 standard deviations\n"
     ]
    }
   ],
   "source": [
    "print(f' dummy gg model RMSE w/ repeat cust only: {np.round(dummy_gg_rmse,2)}, which is equal to {np.round((dummy_gg_rmse/monetary_value_holdout_std),2)} standard deviations')\n",
    "print(f' gg model RMSE w/ repeat cust only: {np.round(gg_1_rmse,2)}, which is equal to {np.round((gg_1_rmse/monetary_value_holdout_std),2)} standard deviations')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Predicting LTV For Return Customers\n",
    "Given the large class imbalence, this prediction will focus only on return customers, those with more than 1 purchase."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_unique_id\n",
       "0000366f3b9a7992bf8c76cfdf3221e2    294.481754\n",
       "0000b849f77a49e4a4ce2b2a4ca5be3f    294.481754\n",
       "0000f46a3911fa3c0805444483337064    294.481754\n",
       "0000f6ccb0745a6a4b88665a16c9f078    294.481754\n",
       "0004aac84e0df4da2b147fca70cf8255    294.481754\n",
       "dtype: float64"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#subset the summary data to just return customers, GG models do not accept a freq of zero\n",
    "return_cust_summary = summary[summary[\"frequency\"]>0]\n",
    "#fit the GG model\n",
    "final_gg = GammaGammaFitter(penalizer_coef = 0.001)\n",
    "final_gg.fit(return_cust_summary['frequency'],\n",
    "        return_cust_summary['monetary_value'])\n",
    "#We can now estimate the average transaction value for all customers:\n",
    "final_gg.conditional_expected_average_profit(\n",
    "        summary['frequency'],\n",
    "        summary['monetary_value']).head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "customer_unique_id\n",
       "004288347e5e88a27ded2bb23747066c     3.196030\n",
       "004b45ec5c64187465168251cd1c9c2f     2.683101\n",
       "00a39521eb40f7012db50455bf083460     1.223831\n",
       "011575986092c30523ecb71ff10cb473     3.019339\n",
       "011b4adcd54683b480c4d841250a987f     5.062672\n",
       "                                      ...    \n",
       "fed519569d16e690df6f89cb99d4e682     3.769938\n",
       "ff03923ad1eb9e32304deb7f9b2a45c9     4.958404\n",
       "ff22e30958c13ffe219db7d711e8f564    14.277565\n",
       "ff8892f7c26aa0446da53d01b18df463     1.882124\n",
       "ff922bdd6bafcdf99cb90d7f39cea5b3     2.979693\n",
       "Name: clv, Length: 2065, dtype: float64"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Now we'll estimate LTV over the next 12 months\n",
    "prediction_period = 12\n",
    "# refit the bgf model to the full return customer data-set\n",
    "bgf_1.fit(summary['frequency'], summary['recency'], summary['T'])\n",
    "\n",
    "ltv_predictions = final_gg.customer_lifetime_value(\n",
    "    bgf_1, #our best bgf model\n",
    "    return_cust_summary['frequency'],\n",
    "    return_cust_summary['recency'],\n",
    "    return_cust_summary['T'],\n",
    "    return_cust_summary['monetary_value'],\n",
    "    time=prediction_period, # months\n",
    "    discount_rate=0.0, # \n",
    "    freq =\"D\"\n",
    ")\n",
    "ltv_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "total revenue from return customers in next 365 days 14072.0\n"
     ]
    }
   ],
   "source": [
    "print(f'total revenue from return customers in next 365 days {np.round(sum(ltv_predictions))}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visuals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "''"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot a histogram of our ltv predictions\n",
    "fig, ax = plt.subplots(figsize=(10,5))\n",
    "ax.hist(ltv_predictions)\n",
    "ax.set_xbound(0,80)\n",
    "ax.set_title(\"Distribution of CLV In Next 12 Months\",fontsize=18)\n",
    "ax.set_ylabel(\"Customer Count\",fontsize=18)\n",
    "ax.set_xlabel(\"Predicted CLV\",fontsize=18)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.savefig(\"visuals/dist_of_clv_in_next_12months.png\");\n",
    ";"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Get a list of predicted CLV for n months in the future\n",
    "prediction_period = 12\n",
    "clv_sum_per_day = []\n",
    "date_range = [*range(1, (prediction_period)+1, 1)]\n",
    "for months in date_range:\n",
    "    ltv_predictions = final_gg.customer_lifetime_value(\n",
    "    bgf_1, #our best bgf model\n",
    "    summary['frequency'],\n",
    "    summary['recency'],\n",
    "    summary['T'],\n",
    "    summary['monetary_value'],\n",
    "    time=months, # months\n",
    "    discount_rate=0.0, # monthly discount rate ~ 12.7% annually\n",
    "    freq =\"D\"\n",
    ")\n",
    "    clv_sum_per_day.append(sum(ltv_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "''"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot a line graph of cumulative CLV over the prediction period\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10,5))\n",
    "ax.plot(date_range,clv_sum_per_day)\n",
    "ax.set_title(\"Cumulative CLV Next 12 Months\",fontsize=18)\n",
    "ax.set_ylabel(\"Cumulative CLV\",fontsize=18)\n",
    "ax.set_xlabel(\"Months in The Future\",fontsize=18)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"visuals/cumulative_clv_future.png\")\n",
    ";"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from lifetimes.plotting import plot_frequency_recency_matrix\n",
    "fig, ax = plt.subplots(figsize=(10,8))\n",
    "fig = plot_frequency_recency_matrix(bgf_1,T=365)\n",
    "ax.set_title(\"Expected Purchases In Next 365 Days\",fontsize=18)\n",
    "fig.set_xlabel(\"Customer's Historical Frequency\",fontsize=18)\n",
    "fig.set_ylabel(\"Customer's Recency\",fontsize=18)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18);\n",
    "plt.savefig(\"visuals/expected_purchases_matrix.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from lifetimes.plotting import plot_probability_alive_matrix\n",
    "fig = plt.figure(figsize=(10,8))\n",
    "fig = plot_probability_alive_matrix(bgf_1)\n",
    "fig.set_title(\"Probability Customer Is Active\",fontsize=18)\n",
    "fig.set_xlabel(\"Customer's Historical Frequency\",fontsize=18)\n",
    "fig.set_ylabel(\"Customer's Recency\",fontsize=18)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    ";\n",
    "plt.savefig(\"visuals/probability_active_matrix.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "r         0.011807\n",
       "alpha    46.464724\n",
       "a         2.650639\n",
       "b         0.374417\n",
       "dtype: float64"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bgf_1.params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot the heterogeneity of Purchase Rate using the parameters of our model\n",
    "import scipy.stats as stats\n",
    "fig, ax_gamma = plt.subplots(figsize = (10, 8))\n",
    "\n",
    "x_gamma = np.linspace(0, 10, 1000)\n",
    "y_gamma = stats.gamma.pdf(x_gamma, a=bgf_1.params_[\"alpha\"], scale=bgf_1.params_[\"r\"])\n",
    "ax_gamma.plot(x_gamma, y_gamma, \"-\")\n",
    "ax_gamma.set_title(f'Gamma: Purchase rate distribution (alpha (α) = {bgf_1.params_[\"alpha\"]:.2f}, scale (r)  = {bgf_1.params_[\"r\"]:.2f})',fontsize=20)\n",
    "ax_gamma.set_xlabel(r\"Purchase Rate \" '($\\lambda$)',fontsize=20)\n",
    "ax_gamma.set_ylabel(r'$P(\\lambda)$',fontsize=20)\n",
    "ax_gamma.set_xbound(0,2)\n",
    "ax_gamma.set_xticks([0,.5,1,1.5,2]);\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.tight_layout();\n",
    "plt.savefig(\"visuals/gamma_purchase_rate_dist.png\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot the heterogeneity of de-activate risk using the parameters of our model\n",
    "#Our customers are not in a healthy position, much of them are highly likely to be de-activated already\n",
    "fig, ax_beta = plt.subplots(figsize = (10, 8))\n",
    "x_beta = np.linspace(0, 1, 100)\n",
    "y_beta = stats.beta.pdf(x_beta, a=bgf_1.params_[\"a\"], b=bgf_1.params_[\"b\"])\n",
    "ax_beta.plot(x_beta, y_beta, \"-\")\n",
    "ax_beta.set_title(f'Beta: De-activate risk distrobution (a = {bgf_1.params_[\"a\"]:.2f}, scale (β)  = {bgf_1.params_[\"b\"]:.2f})',fontsize=20)\n",
    "ax_beta.set_xlabel(\"Probability de-activate after purchase (p)\" ,fontsize=20)\n",
    "ax_beta.set_ylabel('P(p)',fontsize=20)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.tight_layout();\n",
    "plt.savefig(\"visuals/beta_deactivate_risk_dist.png\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax_gamma = plt.subplots(figsize = (10, 8))\n",
    "\n",
    "x_gamma = np.linspace(0, 300, 1000)\n",
    "y_gamma = stats.gamma.pdf(x_gamma, a=gg_1.params_[0], scale=gg_1.params_[2])\n",
    "ax_gamma.plot(x_gamma, y_gamma, \"-\")\n",
    "ax_gamma.set_title(f'Gamma: Avg Trans value dist (p = {gg_1.params_[0]:.2f}, scale (v)  = {gg_1.params_[2]:.2f})',fontsize=20)\n",
    "ax_gamma.set_xlabel(r\"Avg Transaction Value \" '(Z)',fontsize=20)\n",
    "ax_gamma.set_ylabel(r'P (Z)',fontsize=20)\n",
    "plt.xticks(fontsize=18)\n",
    "plt.yticks(fontsize=18)\n",
    "plt.tight_layout();\n",
    "plt.savefig(\"visuals/avg_transaction_value_dist.png\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "## Limitations & Next Steps\n",
    "**Margin**\n",
    "\n",
    "Most CLV formulas hold for product margin as part of the calculation. Multiplying margin by the other factors produces a customer lifetime *profit*. This project didn't have access to margin information thus it wasn't included.\n",
    "\n",
    "Future iterations would include margin if that data became available.\n",
    "\n",
    "**Limited business context applicability**\n",
    "\n",
    "Applicability limited to non-contractual businesses with continuous purchases. There are other business contexts where this modeling approach is not applicable. [Breakdown of Olist business context](https://github.com/zander1268/custorama/blob/main/visuals/business_context.jpg?raw=true).\n",
    "\n",
    "Future iterations of the model could be altered to better suit different business contexts.\n",
    "\n",
    "**Limited number of repeat customers***\n",
    "\n",
    "Because Gamma Gamma model fitters do not accept zero values in freq and monetary values, I had to fit and validate on the limited subset of users with a repeat purchases. Given the class imbalance this is a limited group. Ideally, there would be more data that included a greater number of repeat purchasers to help better validate gamma gamma model and demonstrate the performance of BG/NBD model over the dummy model.\n",
    "\n",
    "## Conclusion\n",
    "\n",
    "This project successfully created a user-friendly CLV prediction engine and validated it's performance on real world transactional data from Olist with minimal error. The app is built on the same data processing and modeling techniques as the validated model but is capable of taking in unstructured and unseen data from new merchants and returning important CLV predictions. The app will empower merchants to make better decisions and delight customers. It opens the door to a number of business analyses; marketing ROI, sales forecasts, company valuation, churn risk, and persona analysis. Olist can quickly roll out this tool to it's e-commerce merchants and enjoy greater commissions and retention rate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (lifetime-env)",
   "language": "python",
   "name": "lifetime-env"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}