{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"import numpy as np\n",
"import boto3\n",
"\n",
"from datetime import timedelta\n",
"\n",
"from sklearn.model_selection import train_test_split\n",
"\n",
"from lifelines.utils import datetimes_to_durations\n",
"from lifelines import CoxPHFitter\n",
"\n",
"from lifetimes import ModifiedBetaGeoFitter, GammaGammaFitter\n",
"from lifetimes.plotting import plot_cumulative_transactions, plot_incremental_transactions, plot_calibration_purchases_vs_holdout_purchases\n",
"from lifetimes.utils import calibration_and_holdout_data\n",
"\n",
"import matplotlib.pyplot as plt\n",
"import seaborn as sns\n",
"\n",
"from utils import woe\n",
"\n",
"%matplotlib inline"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Parameters"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"obs_date = '2019-04-01'\n",
"spline_censor_date = '2020-03-01'\n",
"censor_date = '2020-08-01'\n",
"app_type = ['fresh_premium', 'upgrade']"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"s3 = boto3.client('s3')\n",
"\n",
"def get_matching_s3_keys(bucket, prefix='', suffix=''):\n",
" \"\"\"\n",
" Generate the keys in an S3 bucket.\n",
"\n",
" :param bucket: Name of the S3 bucket.\n",
" :param prefix: Only fetch keys that start with this prefix (optional).\n",
" :param suffix: Only fetch keys that end with this suffix (optional).\n",
" \"\"\"\n",
" kwargs = {'Bucket': bucket, 'Prefix': prefix}\n",
" while True:\n",
" resp = s3.list_objects_v2(**kwargs)\n",
" for obj in resp['Contents']:\n",
" key = obj['Key']\n",
" if key.endswith(suffix):\n",
" yield key\n",
"\n",
" try:\n",
" kwargs['ContinuationToken'] = resp['NextContinuationToken']\n",
" except KeyError:\n",
" break"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"key_ls = []\n",
"for key in get_matching_s3_keys(bucket = 'finaccel-ml-model-data-production', \n",
" prefix = 'zhilal/ltv_final/raw_cohorts', \n",
" suffix = '.parquet'):\n",
" key_ls.append(key)\n",
"main_ls = [i for i in key_ls if 'user_set_' in i]"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"feat_key_ls = []\n",
"for key in get_matching_s3_keys(bucket = 'finaccel-ml-model-data-production', \n",
" prefix = 'zhilal/ltv_final/features', \n",
" suffix = '.parquet'):\n",
" feat_key_ls.append(key)\n",
"feat_ls = feat_key_ls"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"key_ls = []\n",
"for key in get_matching_s3_keys(bucket = 'finaccel-ml-model-data-production', \n",
" prefix = 'zhilal/ltv_final/trx_level', \n",
" suffix = '.parquet'):\n",
" key_ls.append(key)\n",
"trx_ls = key_ls"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"main_ls = sorted(main_ls)[:17]\n",
"feat_ls = sorted(feat_ls)\n",
"trx_ls = sorted(trx_ls)"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"['zhilal/ltv_final/raw_cohorts/user_set_2017-07.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2017-08.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2017-09.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2017-10.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2017-11.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2017-12.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-01.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-02.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-03.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-04.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-05.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-06.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-07.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-08.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-09.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-10.parquet',\n",
" 'zhilal/ltv_final/raw_cohorts/user_set_2018-11.parquet']"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"main_ls"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"['zhilal/ltv_final/features/features_2017-07.parquet',\n",
" 'zhilal/ltv_final/features/features_2017-08.parquet',\n",
" 'zhilal/ltv_final/features/features_2017-09.parquet',\n",
" 'zhilal/ltv_final/features/features_2017-10.parquet',\n",
" 'zhilal/ltv_final/features/features_2017-11.parquet',\n",
" 'zhilal/ltv_final/features/features_2017-12.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-01.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-02.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-03.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-04.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-05.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-06.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-07.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-08.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-09.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-10.parquet',\n",
" 'zhilal/ltv_final/features/features_2018-11.parquet']"
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"feat_ls"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"['zhilal/ltv_final/trx_level/trx_level_2018-07.parquet',\n",
" 'zhilal/ltv_final/trx_level/trx_level_2018-08.parquet',\n",
" 'zhilal/ltv_final/trx_level/trx_level_2018-09.parquet',\n",
" 'zhilal/ltv_final/trx_level/trx_level_2018-10.parquet',\n",
" 'zhilal/ltv_final/trx_level/trx_level_2018-11.parquet']"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"trx_ls"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# 0. Get all the data"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Users"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [],
"source": [
"rel_cols = [\n",
" 'user_id'\n",
" , 'set_password_timestamp'\n",
" , 'is_approved_timestamp'\n",
" , 'first_dpd_91_dt'\n",
" , 'first_trx_date'\n",
" , 'application_type'\n",
" , 'user_status'\n",
"]\n",
"\n",
"main_df = pd.DataFrame()\n",
"for key in main_ls:\n",
" temp_df = pd.read_parquet(r's3://finaccel-ml-model-data-production/'+key, columns=rel_cols)\n",
" \n",
" # Make sure there is no user with user_status = 3 (Rejected) or 11 (blacklisted) \n",
" temp_df = temp_df.query(\"user_status in (2,5,9,7)\")\n",
" \n",
" temp_df['application_type_code'] = temp_df['application_type']\n",
" temp_df['application_type'] = np.where(temp_df['application_type'].isin([150, 160, 170, 180]),\n",
" 'upgrade', np.where(temp_df['application_type'].isin([100, 110, 120]), 'fresh_premium', 'basic'))\n",
" temp_df['application_type'] = temp_df['application_type'].astype('str')\n",
" \n",
" # determine app_type before obs_date\n",
" temp_df['app_type_b4_obsdt'] = np.where(temp_df['is_approved_timestamp'] >= pd.to_datetime(obs_date),\n",
" 'basic', temp_df['application_type'])\n",
" # Select specific app_type only\n",
" temp_df = temp_df[temp_df['app_type_b4_obsdt'].isin(app_type)].reset_index(drop=True) \n",
" \n",
" # remove users who NPL before obs_date\n",
" temp_df = temp_df[~(temp_df['first_dpd_91_dt'] < \n",
" pd.to_datetime(obs_date))].reset_index(drop=True)\n",
"\n",
" # remove users who have not made\n",
" # 1st trx before obs_date\n",
" temp_df = temp_df[~(temp_df['first_trx_date'] > pd.to_datetime(obs_date)) & \n",
" temp_df['first_trx_date'].notna()]\n",
" \n",
" main_df = pd.concat([main_df, temp_df]).reset_index(drop=True)"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(108965, 9)"
]
},
"execution_count": 12,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"main_df.shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Features"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {
"scrolled": true
},
"outputs": [],
"source": [
"temp_ls = []\n",
"for feat_key in feat_ls:\n",
" temp_df = pd.read_parquet('s3://finaccel-ml-model-data-production/'+feat_key)\n",
" temp_ls.append(temp_df)\n",
" \n",
"feat_df = pd.concat(temp_ls).reset_index(drop=True)\n",
"feat_df = feat_df.drop_duplicates('user_id', 'first').reset_index(drop=True)"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(357579, 415)"
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"feat_df.shape"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [],
"source": [
"main_df = main_df.merge(feat_df, how='left', on='user_id')"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(108965, 423)"
]
},
"execution_count": 16,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"main_df.shape"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [],
"source": [
"del temp_ls, temp_df, feat_df"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Transactions"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [],
"source": [
"temp_ls = []\n",
"for key in trx_ls:\n",
" temp_df = pd.read_parquet('s3://finaccel-ml-model-data-production/'+key)\n",
" temp_ls.append(temp_df)\n",
" \n",
"trx_df = pd.concat(temp_ls).reset_index(drop=True)\n",
"\n",
"del temp_ls, temp_df"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"trx_df['app_type_b4_obsdt'] = np.where(trx_df['is_approved_timestamp'] > pd.to_datetime(obs_date),\n",
" 'basic', trx_df['application_type'])\n",
"trx_df['app_type'] = np.where(trx_df['app_type_b4_obsdt'] == 'basic', 'basic', 'premium')\n",
"trx_df = trx_df[trx_df['app_type_b4_obsdt'].isin(app_type)].reset_index(drop=True)"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(2428041, 29)"
]
},
"execution_count": 20,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"trx_df.shape"
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"54199"
]
},
"execution_count": 21,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"trx_df['user_id'].nunique()"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [],
"source": [
"trx_df['trx_dt'] = trx_df['transaction_date'].dt.date.astype('datetime64')\n",
"\n",
"trx_df = trx_df.query(f'trx_dt < \"{censor_date}\"')"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(2281764, 30)"
]
},
"execution_count": 23,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"trx_df.shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Adjust the LTV components\n",
"- use the LTV component coefficients"
]
},
{
"cell_type": "code",
"execution_count": 26,
"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>app_type</th>\n",
" <th>payment_type</th>\n",
" <th>prop_collections_cost</th>\n",
" <th>prop_latefee</th>\n",
" <th>prop_lateinterest</th>\n",
" <th>prop_lender_interest_fees</th>\n",
" <th>prop_reg_interest</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>basic</td>\n",
" <td>30_days</td>\n",
" <td>0.0058</td>\n",
" <td>0.0300</td>\n",
" <td>0.0295</td>\n",
" <td>0.0163</td>\n",
" <td>0.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>premium</td>\n",
" <td>12_months</td>\n",
" <td>0.0325</td>\n",
" <td>0.0341</td>\n",
" <td>0.0359</td>\n",
" <td>0.0935</td>\n",
" <td>0.2402</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>premium</td>\n",
" <td>30_days</td>\n",
" <td>0.0058</td>\n",
" <td>0.0600</td>\n",
" <td>0.0400</td>\n",
" <td>0.0186</td>\n",
" <td>0.0000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>premium</td>\n",
" <td>3_months</td>\n",
" <td>0.0104</td>\n",
" <td>0.0321</td>\n",
" <td>0.0305</td>\n",
" <td>0.0290</td>\n",
" <td>0.0704</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>premium</td>\n",
" <td>6_months</td>\n",
" <td>0.0177</td>\n",
" <td>0.0394</td>\n",
" <td>0.0376</td>\n",
" <td>0.0506</td>\n",
" <td>0.1252</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" app_type payment_type prop_collections_cost prop_latefee \\\n",
"0 basic 30_days 0.0058 0.0300 \n",
"1 premium 12_months 0.0325 0.0341 \n",
"2 premium 30_days 0.0058 0.0600 \n",
"3 premium 3_months 0.0104 0.0321 \n",
"4 premium 6_months 0.0177 0.0394 \n",
"\n",
" prop_lateinterest prop_lender_interest_fees prop_reg_interest \n",
"0 0.0295 0.0163 0.0000 \n",
"1 0.0359 0.0935 0.2402 \n",
"2 0.0400 0.0186 0.0000 \n",
"3 0.0305 0.0290 0.0704 \n",
"4 0.0376 0.0506 0.1252 "
]
},
"execution_count": 26,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"adj_df = pd.read_parquet('s3://finaccel-ml-model-data-production/zhilal/ltv_final/ltv_components_proportions_20201110.parquet')\n",
"adj_df"
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [],
"source": [
"trx_df = trx_df.merge(adj_df, how='left', on=['app_type', 'payment_type'])"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [],
"source": [
"payment_type_dict = {'30_days':30,\n",
" '3_months':90,\n",
" '6_months':180,\n",
" '12_months':360}\n",
"trx_df['payment_type_int'] = trx_df['payment_type'].map(payment_type_dict)"
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {},
"outputs": [],
"source": [
"comp_ls = [\n",
" 'reg_interest',\n",
" 'lateinterest',\n",
" 'latefee',\n",
" 'lender_interest_fees',\n",
" 'collections_cost'\n",
"]"
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [],
"source": [
"def recalc_ltv_comp(df, timebound, components):\n",
" df['trx_dob'] = (pd.to_datetime(timebound) - df['transaction_date']).dt.days\n",
" df['flag_mature'] = np.where(df['trx_dob'] >= df['payment_type_int'], 1, 0)\n",
" display(df['flag_mature'].value_counts())\n",
"\n",
" mature_df = df[df['flag_mature'] == 1].reset_index(drop=True)\n",
" print('mature df shape:', mature_df.shape)\n",
"\n",
" immature_df = df[df['flag_mature'] == 0].reset_index(drop=True)\n",
" print('immature df shape:', immature_df.shape)\n",
"\n",
" for comp in components:\n",
" immature_df[comp] = immature_df['orig_principal'] * immature_df['prop_'+comp]\n",
"\n",
" fin = pd.concat([mature_df, immature_df]).reset_index(drop=True)\n",
" print('final dataset size:', fin.shape)\n",
" \n",
" return fin"
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(2281764, 36)"
]
},
"execution_count": 31,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"trx_df.shape"
]
},
{
"cell_type": "code",
"execution_count": 32,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"51498"
]
},
"execution_count": 32,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"trx_df['user_id'].nunique()"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Dataset size: (881541, 36)\n",
"\n"
]
},
{
"data": {
"text/plain": [
"1 694789\n",
"0 186752\n",
"Name: flag_mature, dtype: int64"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"mature df shape: (694789, 38)\n",
"immature df shape: (186752, 38)\n",
"final dataset size: (881541, 38)\n"
]
}
],
"source": [
"cal_df = trx_df[trx_df['transaction_date'] < pd.to_datetime(obs_date)].reset_index(drop=True)\n",
"print('Dataset size:', cal_df.shape)\n",
"print()\n",
"cal_fin = recalc_ltv_comp(cal_df, obs_date, comp_ls)"
]
},
{
"cell_type": "code",
"execution_count": 34,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Dataset size: (1400223, 36)\n",
"\n"
]
},
{
"data": {
"text/plain": [
"1 1288914\n",
"0 111309\n",
"Name: flag_mature, dtype: int64"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"mature df shape: (1288914, 38)\n",
"immature df shape: (111309, 38)\n",
"final dataset size: (1400223, 38)\n"
]
}
],
"source": [
"hld_df = trx_df[(trx_df['transaction_date'] >= pd.to_datetime(obs_date)) &\n",
" (trx_df['transaction_date'] < pd.to_datetime(censor_date))].reset_index(drop=True)\n",
"print('Dataset size:', hld_df.shape)\n",
"print()\n",
"hld_fin = recalc_ltv_comp(hld_df, censor_date, comp_ls)"
]
},
{
"cell_type": "code",
"execution_count": 35,
"metadata": {},
"outputs": [],
"source": [
"trx_df = pd.concat([cal_fin, hld_fin]).reset_index(drop=True)"
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(2281764, 38)"
]
},
"execution_count": 36,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"trx_df.shape"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"51498"
]
},
"execution_count": 37,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"trx_df['user_id'].nunique()"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [],
"source": [
"trx_df['rev'] = trx_df['merchant_fee'] + \\\n",
" trx_df['pl_fee'] + \\\n",
" trx_df['processing_p30_fee'] + \\\n",
" trx_df['travel_fee'] + \\\n",
" trx_df['biller_fee'] + \\\n",
" trx_df['reg_interest'] + \\\n",
" trx_df['latefee'] + \\\n",
" trx_df['lateinterest'] - \\\n",
" trx_df['lender_interest_fees'] - \\\n",
" trx_df['collections_cost'] - \\\n",
" trx_df['discount_cost']"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Prepare the Framework Calibration (train) and Holdout (test) Data"
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {},
"outputs": [],
"source": [
"trx_df['ltv_calib'] = np.where(trx_df['trx_dt'] < obs_date, trx_df['rev'], 0)\n",
"\n",
"trx_df['act_rev_holdout'] = np.where(trx_df['trx_dt'] > obs_date, trx_df['rev'], 0)\n",
"\n",
"trx_df['trx_dpd_91_dt'] = trx_df['trx_dpd_91_dt'].astype('datetime64[s]')\n",
"trx_df['act_npl_holdout'] = np.where((trx_df['trx_dpd_91_dt'] > pd.to_datetime(obs_date)) & \n",
" (trx_df['trx_dpd_91_dt'] < pd.to_datetime(censor_date))\n",
" , trx_df['npl_expenses'], 0)\n",
"\n",
"trx_df['act_ltv_holdout'] = trx_df['act_rev_holdout'] - trx_df['act_npl_holdout']"
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {},
"outputs": [],
"source": [
"# this is the summary of ltv data\n",
"\n",
"main_ltv_df = trx_df.groupby('user_id').agg({'ltv_calib':sum,\n",
" 'act_ltv_holdout':sum, \n",
" 'act_npl_holdout':sum, \n",
" 'act_rev_holdout':sum\n",
" }).reset_index()\n",
"\n",
"main_ltv_df = main_ltv_df.merge(main_df[['user_id','snapshot_os_principal']],\n",
" how = 'right', on ='user_id').fillna(0)"
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>user_id</th>\n",
" <th>ltv_calib</th>\n",
" <th>act_ltv_holdout</th>\n",
" <th>act_npl_holdout</th>\n",
" <th>act_rev_holdout</th>\n",
" <th>snapshot_os_principal</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1100743</td>\n",
" <td>1.151064e+06</td>\n",
" <td>2.486893e+06</td>\n",
" <td>0.0</td>\n",
" <td>2.486893e+06</td>\n",
" <td>3920210.90</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1146340</td>\n",
" <td>2.573292e+05</td>\n",
" <td>5.464701e+05</td>\n",
" <td>0.0</td>\n",
" <td>5.464701e+05</td>\n",
" <td>1180471.34</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>1189034</td>\n",
" <td>1.654805e+06</td>\n",
" <td>3.830105e+06</td>\n",
" <td>0.0</td>\n",
" <td>3.830105e+06</td>\n",
" <td>3341407.01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>1246458</td>\n",
" <td>2.239381e+06</td>\n",
" <td>2.970454e+06</td>\n",
" <td>0.0</td>\n",
" <td>2.970454e+06</td>\n",
" <td>8119529.72</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>1344165</td>\n",
" <td>1.153228e+06</td>\n",
" <td>1.300245e+06</td>\n",
" <td>0.0</td>\n",
" <td>1.300245e+06</td>\n",
" <td>5713492.71</td>\n",
" </tr>\n",
" <tr>\n",
" <th>...</th>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>108960</th>\n",
" <td>3786609</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.0</td>\n",
" <td>0.000000e+00</td>\n",
" <td>4417726.83</td>\n",
" </tr>\n",
" <tr>\n",
" <th>108961</th>\n",
" <td>3818225</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.0</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>108962</th>\n",
" <td>3788341</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.0</td>\n",
" <td>0.000000e+00</td>\n",
" <td>6817124.54</td>\n",
" </tr>\n",
" <tr>\n",
" <th>108963</th>\n",
" <td>3566616</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.0</td>\n",
" <td>0.000000e+00</td>\n",
" <td>22285673.31</td>\n",
" </tr>\n",
" <tr>\n",
" <th>108964</th>\n",
" <td>5015875</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.0</td>\n",
" <td>0.000000e+00</td>\n",
" <td>0.00</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>108965 rows × 6 columns</p>\n",
"</div>"
],
"text/plain": [
" user_id ltv_calib act_ltv_holdout act_npl_holdout \\\n",
"0 1100743 1.151064e+06 2.486893e+06 0.0 \n",
"1 1146340 2.573292e+05 5.464701e+05 0.0 \n",
"2 1189034 1.654805e+06 3.830105e+06 0.0 \n",
"3 1246458 2.239381e+06 2.970454e+06 0.0 \n",
"4 1344165 1.153228e+06 1.300245e+06 0.0 \n",
"... ... ... ... ... \n",
"108960 3786609 0.000000e+00 0.000000e+00 0.0 \n",
"108961 3818225 0.000000e+00 0.000000e+00 0.0 \n",
"108962 3788341 0.000000e+00 0.000000e+00 0.0 \n",
"108963 3566616 0.000000e+00 0.000000e+00 0.0 \n",
"108964 5015875 0.000000e+00 0.000000e+00 0.0 \n",
"\n",
" act_rev_holdout snapshot_os_principal \n",
"0 2.486893e+06 3920210.90 \n",
"1 5.464701e+05 1180471.34 \n",
"2 3.830105e+06 3341407.01 \n",
"3 2.970454e+06 8119529.72 \n",
"4 1.300245e+06 5713492.71 \n",
"... ... ... \n",
"108960 0.000000e+00 4417726.83 \n",
"108961 0.000000e+00 0.00 \n",
"108962 0.000000e+00 6817124.54 \n",
"108963 0.000000e+00 22285673.31 \n",
"108964 0.000000e+00 0.00 \n",
"\n",
"[108965 rows x 6 columns]"
]
},
"execution_count": 42,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"main_ltv_df"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"check the data"
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(108965, 423) (2281764, 43) (108965, 6)\n"
]
}
],
"source": [
"print(main_df.shape, trx_df.shape, main_ltv_df.shape)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# 1. Survival Model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Prepare the \"duration\" and \"flag_churn\" variables"
]
},
{
"cell_type": "code",
"execution_count": 44,
"metadata": {
"scrolled": true
},
"outputs": [],
"source": [
"# censor NPL events that happen after censor date\n",
"\n",
"main_df['first_dpd_91_censor'] = np.where(main_df['first_dpd_91_dt'] < pd.to_datetime(spline_censor_date), \n",
" main_df['first_dpd_91_dt'], None)\n",
"main_df['first_dpd_91_censor'] = main_df['first_dpd_91_censor'].astype('datetime64')\n",
"\n",
"# also set the first_dpd_91 to 1st of month\n",
"main_df['first_dpd_91_censor_bom'] = main_df['first_dpd_91_censor'] + pd.offsets.MonthEnd(0) - pd.offsets.MonthBegin(normalize=True)\n",
"\n",
"main_df['t0'] = pd.to_datetime(obs_date)"
]
},
{
"cell_type": "code",
"execution_count": 45,
"metadata": {},
"outputs": [],
"source": [
"main_df['duration_t0'] = np.where(main_df['first_dpd_91_censor'].isna(), 11, \n",
" main_df['first_dpd_91_censor'].dt.year*12 + \\\n",
" main_df['first_dpd_91_censor'].dt.month - \\\n",
" 2019*12 - 3)"
]
},
{
"cell_type": "code",
"execution_count": 46,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1.0 1603\n",
"2.0 1614\n",
"3.0 1420\n",
"4.0 1238\n",
"5.0 1030\n",
"6.0 997\n",
"7.0 1014\n",
"8.0 1135\n",
"9.0 1107\n",
"10.0 1022\n",
"11.0 96785\n",
"Name: duration_t0, dtype: int64"
]
},
"execution_count": 46,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"main_df['duration_t0'].value_counts().sort_index()"
]
},
{
"cell_type": "code",
"execution_count": 47,
"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 tr th {\n",
" text-align: left;\n",
" }\n",
"\n",
" .dataframe thead tr:last-of-type th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr>\n",
" <th></th>\n",
" <th colspan=\"2\" halign=\"left\">first_dpd_91_censor_bom</th>\n",
" </tr>\n",
" <tr>\n",
" <th></th>\n",
" <th>max</th>\n",
" <th>min</th>\n",
" </tr>\n",
" <tr>\n",
" <th>duration_t0</th>\n",
" <th></th>\n",
" <th></th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>1.0</th>\n",
" <td>2019-04-01</td>\n",
" <td>2019-04-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2.0</th>\n",
" <td>2019-05-01</td>\n",
" <td>2019-05-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3.0</th>\n",
" <td>2019-06-01</td>\n",
" <td>2019-06-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4.0</th>\n",
" <td>2019-07-01</td>\n",
" <td>2019-07-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5.0</th>\n",
" <td>2019-08-01</td>\n",
" <td>2019-08-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6.0</th>\n",
" <td>2019-09-01</td>\n",
" <td>2019-09-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7.0</th>\n",
" <td>2019-10-01</td>\n",
" <td>2019-10-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8.0</th>\n",
" <td>2019-11-01</td>\n",
" <td>2019-11-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9.0</th>\n",
" <td>2019-12-01</td>\n",
" <td>2019-12-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10.0</th>\n",
" <td>2020-01-01</td>\n",
" <td>2020-01-01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11.0</th>\n",
" <td>2020-02-01</td>\n",
" <td>2020-02-01</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" first_dpd_91_censor_bom \n",
" max min\n",
"duration_t0 \n",
"1.0 2019-04-01 2019-04-01\n",
"2.0 2019-05-01 2019-05-01\n",
"3.0 2019-06-01 2019-06-01\n",
"4.0 2019-07-01 2019-07-01\n",
"5.0 2019-08-01 2019-08-01\n",
"6.0 2019-09-01 2019-09-01\n",
"7.0 2019-10-01 2019-10-01\n",
"8.0 2019-11-01 2019-11-01\n",
"9.0 2019-12-01 2019-12-01\n",
"10.0 2020-01-01 2020-01-01\n",
"11.0 2020-02-01 2020-02-01"
]
},
"execution_count": 47,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"main_df.groupby('duration_t0').agg({'first_dpd_91_censor_bom':['max', 'min']})"
]
},
{
"cell_type": "code",
"execution_count": 48,
"metadata": {},
"outputs": [],
"source": [
"main_df['flag_churn'] = main_df['first_dpd_91_censor'].notna()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Minor Feature Transformation"
]
},
{
"cell_type": "code",
"execution_count": 52,
"metadata": {},
"outputs": [],
"source": [
"main_df['util_latest'] = np.where(main_df['util_latest'] > 1, 1, main_df['util_latest'])\n",
"\n",
"education_mapping = {'pre_high_school': 1, \n",
" 'high_school': 2, \n",
" 'diploma': 3,\n",
" 'undergraduate': 4, \n",
" 'postgraduate': 5}\n",
"main_df['de_education'] = main_df['de_education'].map(education_mapping)\n",
"main_df['snapshot_days_premium'] = main_df['snapshot_days_premium'].fillna(120)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Train - Validation - Test Split"
]
},
{
"cell_type": "code",
"execution_count": 53,
"metadata": {},
"outputs": [],
"source": [
"train_val_temp, test_df = train_test_split(main_df, test_size=0.1, random_state=99)\n",
"train_df, val_df = train_test_split(train_val_temp, test_size=0.1, random_state=99)"
]
},
{
"cell_type": "code",
"execution_count": 54,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"train size: (88261, 428)\n",
"val size: (9807, 428)\n",
"test size: (10897, 428)\n"
]
}
],
"source": [
"print('train size:', train_temp.shape)\n",
"print('val size:', val_df.shape)\n",
"print('test size:', test_df.shape)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Selected Features"
]
},
{
"cell_type": "code",
"execution_count": 56,
"metadata": {},
"outputs": [],
"source": [
"cont_feats = [\n",
" 'ap_co_plo',\n",
" 'delin_dpd_max_90',\n",
" 'rep_count_30',\n",
" 'snapshot_current_dpd',\n",
" 'snapshot_days_premium',\n",
" 'snapshot_os_amount',\n",
" 'trx_active_amt_avg',\n",
" 'trx_denied_count_120',\n",
" 'trx_merch_count_dist_60',\n",
" 'trx_pl_count_90',\n",
" 'trx_suc_amt_sum_30',\n",
" 'util_latest',\n",
"]\n",
"\n",
"disc_feats = [\n",
" 'de_education'\n",
"]\n",
"\n",
"rel_model_cols = [\n",
" 'user_id',\n",
" 'duration_t0',\n",
" 'flag_churn',\n",
" 'app_type_b4_obsdt'\n",
"]\n",
"\n",
"feat_dict = {**{x:'c' for x in cont_feats},\n",
" **{x:'d' for x in disc_feats},\n",
" }"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### WoE Feature Transformation"
]
},
{
"cell_type": "code",
"execution_count": 58,
"metadata": {},
"outputs": [],
"source": [
"def woe_transform(train, test, feature, qnt_num, v_type, target_var):\n",
" print(feature)\n",
" if v_type == 'c':\n",
" try:\n",
" _, bins_man = pd.qcut(train[feature], qnt_num, retbins=True, duplicates='drop')\n",
" except:\n",
" return np.nan\n",
" wo_c = woe.WoE(qnt_num=qnt_num, v_type=v_type, bins=bins_man)\n",
" elif v_type == 'd':\n",
" wo_c = woe.WoE(v_type=v_type)\n",
"\n",
" wo_temp = wo_c.fit(train[feature], train[target_var])\n",
" tr_df = wo_temp.transform(train[feature]).rename(columns={'woe':'std_'+feature})[['std_'+feature]]\n",
" te_df = wo_temp.transform(test[feature]).rename(columns={'woe':'std_'+feature})[['std_'+feature]]\n",
" return wo_temp, tr_df, te_df"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"train_ls = []\n",
"val_ls = []\n",
"test_ls = []\n",
"for feature, d_type in feat_dict.items():\n",
" info, temp_tr, temp_val = woe_transform(train_df, val_df, feature, 10, d_type, 'flag_churn')\n",
" info, temp_tr, temp_te = woe_transform(train_df, test_df, feature, 10, d_type, 'flag_churn')\n",
" train_ls.append(temp_tr)\n",
" val_ls.append(temp_val)\n",
" test_ls.append(temp_te)\n",
" print(info.iv)\n",
" display(info.bins)\n",
" print()\n",
" \n",
"std_train = pd.concat(train_ls, axis=1)\n",
"std_val = pd.concat(val_ls, axis=1)\n",
"std_test = pd.concat(test_ls, axis=1)"
]
},
{
"cell_type": "code",
"execution_count": 64,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"train (13239, 428)\n",
"val (9807, 428)\n",
"test (10897, 428)\n"
]
}
],
"source": [
"train_df = train_df.merge(std_train, how='left', left_index=True, right_index=True)\n",
"val_df = val_df.merge(std_val, how='left', left_index=True, right_index=True)\n",
"test_df = test_df.merge(std_test, how='left', left_index=True, right_index=True)\n",
"\n",
"print('train', train_df.shape)\n",
"print('val', val_df.shape)\n",
"print('test', test_df.shape)\n",
"\n",
"std_cont_feats = ['std_'+feat for feat in cont_feats]\n",
"std_disc_feats = ['std_'+feat for feat in disc_feats]\n",
"std_features = std_cont_feats + std_disc_feats"
]
},
{
"cell_type": "code",
"execution_count": 65,
"metadata": {},
"outputs": [],
"source": [
"final_feats = sorted(std_features)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Spline Cox Regression Model"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"scrolled": true
},
"outputs": [],
"source": [
"cph_spline = CoxPHFitter(baseline_estimation_method=\"spline\", n_baseline_knots=2)\n",
"cph_spline.fit(train_df[['duration_t0','flag_churn']+final_feats], 'duration_t0', 'flag_churn')"
]
},
{
"cell_type": "code",
"execution_count": 64,
"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>baseline hazard</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>1.0</th>\n",
" <td>0.003548</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2.0</th>\n",
" <td>0.005672</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3.0</th>\n",
" <td>0.006858</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4.0</th>\n",
" <td>0.007417</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5.0</th>\n",
" <td>0.007569</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6.0</th>\n",
" <td>0.007557</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7.0</th>\n",
" <td>0.007548</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8.0</th>\n",
" <td>0.007565</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9.0</th>\n",
" <td>0.007613</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10.0</th>\n",
" <td>0.007691</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11.0</th>\n",
" <td>0.007797</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" baseline hazard\n",
"1.0 0.003548\n",
"2.0 0.005672\n",
"3.0 0.006858\n",
"4.0 0.007417\n",
"5.0 0.007569\n",
"6.0 0.007557\n",
"7.0 0.007548\n",
"8.0 0.007565\n",
"9.0 0.007613\n",
"10.0 0.007691\n",
"11.0 0.007797"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"(0.0, 0.009)"
]
},
"execution_count": 64,
"metadata": {},
"output_type": "execute_result"
},
{
"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": [
"# months level baseline hazard\n",
"display(cph_spline.baseline_hazard_)\n",
"plt.plot(cph_spline.baseline_hazard_)\n",
"plt.ylim([0,0.009])"
]
},
{
"cell_type": "code",
"execution_count": 65,
"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>baseline hazard</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>1.0</th>\n",
" <td>0.003548</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2.0</th>\n",
" <td>0.005672</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3.0</th>\n",
" <td>0.006858</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4.0</th>\n",
" <td>0.007417</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5.0</th>\n",
" <td>0.007569</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6.0</th>\n",
" <td>0.007557</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7.0</th>\n",
" <td>0.007548</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8.0</th>\n",
" <td>0.007565</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9.0</th>\n",
" <td>0.007613</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10.0</th>\n",
" <td>0.007691</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11.0</th>\n",
" <td>0.007797</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12.0</th>\n",
" <td>0.007912</td>\n",
" </tr>\n",
" <tr>\n",
" <th>13.0</th>\n",
" <td>0.008020</td>\n",
" </tr>\n",
" <tr>\n",
" <th>14.0</th>\n",
" <td>0.008120</td>\n",
" </tr>\n",
" <tr>\n",
" <th>15.0</th>\n",
" <td>0.008216</td>\n",
" </tr>\n",
" <tr>\n",
" <th>16.0</th>\n",
" <td>0.008305</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" baseline hazard\n",
"1.0 0.003548\n",
"2.0 0.005672\n",
"3.0 0.006858\n",
"4.0 0.007417\n",
"5.0 0.007569\n",
"6.0 0.007557\n",
"7.0 0.007548\n",
"8.0 0.007565\n",
"9.0 0.007613\n",
"10.0 0.007691\n",
"11.0 0.007797\n",
"12.0 0.007912\n",
"13.0 0.008020\n",
"14.0 0.008120\n",
"15.0 0.008216\n",
"16.0 0.008305"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"[<matplotlib.lines.Line2D at 0x7f7ea8ef6ee0>]"
]
},
"execution_count": 65,
"metadata": {},
"output_type": "execute_result"
},
{
"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": [
"display(cph_spline.baseline_hazard_at_times([i for i in range(1,17,1)]))\n",
"plt.plot(cph_spline.baseline_hazard_at_times([i for i in range(1,17,1)]))"
]
},
{
"cell_type": "code",
"execution_count": 66,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/ec2-user/anaconda3/envs/ltv/lib/python3.8/site-packages/lifelines/fitters/__init__.py:2293: ApproximationWarning: Approximating using `predict_survival_function`. To increase accuracy, try using or increasing the resolution of the timeline kwarg in `.fit(..., timeline=timeline)`.\n",
"\n",
" warnings.warn(\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Train Concordance: 0.7115967281419112\n",
"Val Concordance: 0.7246996721629343\n",
"Test Concordance: 0.723204170938536\n"
]
}
],
"source": [
"print('Train Concordance:',cph_spline.score(train_df, scoring_method='concordance_index'))\n",
"print('Val Concordance:',cph_spline.score(val_df, scoring_method='concordance_index'))\n",
"print('Test Concordance:',cph_spline.score(test_df, scoring_method='concordance_index'))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# 2. MBG NBD Model"
]
},
{
"cell_type": "code",
"execution_count": 69,
"metadata": {},
"outputs": [],
"source": [
"summary_cal_holdout = calibration_and_holdout_data(trx_df, \n",
" customer_id_col = 'user_id', \n",
" datetime_col = 'trx_dt', \n",
" monetary_value_col = 'rev',\n",
" freq = 'D', \n",
" calibration_period_end = pd.to_datetime(obs_date),\n",
" observation_period_end = pd.to_datetime(censor_date))\n",
"\n",
"summary_cal_holdout = summary_cal_holdout.reset_index()\n",
"\n",
"# summary_cal_holdout.to_parquet(\"summary_cal_holdout.parquet\", compression='gzip')"
]
},
{
"cell_type": "code",
"execution_count": 70,
"metadata": {},
"outputs": [],
"source": [
"# summary_cal_holdout = pd.read_parquet(\"summary_cal_holdout.parquet\")"
]
},
{
"cell_type": "code",
"execution_count": 71,
"metadata": {},
"outputs": [],
"source": [
"uid_tr = tuple(train_df['user_id'])\n",
"uid_tr_good = tuple(train_df.query('first_dpd_91_dt != first_dpd_91_dt')['user_id'])\n",
"\n",
"uid_test = tuple(test_df['user_id'])\n",
"uid_test_good = tuple(test_df.query('first_dpd_91_dt != first_dpd_91_dt')['user_id'])\n",
"uid_test_bad = tuple(test_df.query('first_dpd_91_dt == first_dpd_91_dt')['user_id'])"
]
},
{
"cell_type": "code",
"execution_count": 72,
"metadata": {},
"outputs": [],
"source": [
"summary_cal_holdout_tr = summary_cal_holdout[summary_cal_holdout.user_id.isin(uid_tr)]\n",
"summary_cal_holdout_tr_good = summary_cal_holdout[summary_cal_holdout.user_id.isin(uid_tr_good)]\n",
"\n",
"summary_cal_holdout_test = summary_cal_holdout[summary_cal_holdout.user_id.isin(uid_test)]\n",
"summary_cal_holdout_test_good = summary_cal_holdout[summary_cal_holdout.user_id.isin(uid_test_good)]\n",
"summary_cal_holdout_test_bad = summary_cal_holdout[summary_cal_holdout.user_id.isin(uid_test_bad)]"
]
},
{
"cell_type": "code",
"execution_count": 73,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Optimization terminated successfully.\n",
" Current function value: -33.581301\n",
" Iterations: 27\n",
" Function evaluations: 30\n",
" Gradient evaluations: 30\n"
]
},
{
"data": {
"text/plain": [
"<lifetimes.ModifiedBetaGeoFitter: fitted with 27453 subjects, a: 0.02, alpha: 13.23, b: 0.47, r: 1.10>"
]
},
"execution_count": 73,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"mbgnbd = ModifiedBetaGeoFitter(penalizer_coef=0.01)\n",
"mbgnbd.fit(summary_cal_holdout_tr_good['frequency_cal'], \n",
" summary_cal_holdout_tr_good['recency_cal'], \n",
" summary_cal_holdout_tr_good['T_cal'],\n",
" verbose=True,)"
]
},
{
"cell_type": "code",
"execution_count": 74,
"metadata": {},
"outputs": [],
"source": [
"t = (pd.to_datetime(censor_date) - trx_df['trx_dt'].min()).days\n",
"t_cal = (pd.to_datetime(obs_date) - trx_df['trx_dt'].min()).days"
]
},
{
"cell_type": "code",
"execution_count": 75,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"488"
]
},
"execution_count": 75,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"summary_cal_holdout['duration_holdout'].max().astype(int)"
]
},
{
"cell_type": "code",
"execution_count": 76,
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"all\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"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"only good\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"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"only bad\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"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"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": [
"# train using all\n",
"holdout_dur_max = summary_cal_holdout['duration_holdout'].max().astype(int)\n",
"\n",
"print('all')\n",
"plot_cumulative_transactions(mbgnbd, trx_df.query(f'user_id in {uid_test}'), 'trx_dt', 'user_id', t, t_cal, freq='D')\n",
"plt.show()\n",
"\n",
"plot_incremental_transactions(mbgnbd, trx_df.query(f'user_id in {uid_test}'), 'trx_dt', 'user_id', t, t_cal, freq='D')\n",
"plt.show()\n",
"\n",
"plot_calibration_purchases_vs_holdout_purchases(mbgnbd, summary_cal_holdout_test, n=holdout_dur_max)\n",
"sns.despine();\n",
"plt.show()\n",
"\n",
"print('only good')\n",
"plot_cumulative_transactions(mbgnbd, trx_df.query(f'user_id in {uid_test_good}'), 'trx_dt', 'user_id', t, t_cal, freq='D')\n",
"plt.show()\n",
"\n",
"plot_incremental_transactions(mbgnbd, trx_df.query(f'user_id in {uid_test_good}'), 'trx_dt', 'user_id', t, t_cal, freq='D')\n",
"plt.show()\n",
"\n",
"plot_calibration_purchases_vs_holdout_purchases(mbgnbd, summary_cal_holdout_test_good, n=holdout_dur_max)\n",
"sns.despine();\n",
"plt.show()\n",
"\n",
"\n",
"print('only bad')\n",
"plot_cumulative_transactions(mbgnbd, trx_df.query(f'user_id in {uid_test_bad}'), 'trx_dt', 'user_id', t, t_cal, freq='D')\n",
"plt.show()\n",
"\n",
"plot_incremental_transactions(mbgnbd, trx_df.query(f'user_id in {uid_test_bad}'), 'trx_dt', 'user_id', t, t_cal, freq='D')\n",
"plt.show()\n",
"\n",
"plot_calibration_purchases_vs_holdout_purchases(mbgnbd, summary_cal_holdout_test_bad, n=holdout_dur_max)\n",
"sns.despine();\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 77,
"metadata": {},
"outputs": [],
"source": [
"# mbgnbd.save_model('mbgnbd_premium.p')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# 3. Gamma Gamma"
]
},
{
"cell_type": "code",
"execution_count": 78,
"metadata": {},
"outputs": [],
"source": [
"returning_customers_summary = summary_cal_holdout_tr_good.query('frequency_cal >0 and monetary_value_cal > -50000')\n",
"returning_customers_summary['monetary_value_cal'] = \\\n",
" np.where(returning_customers_summary['monetary_value_cal'] <= 0, 1, \n",
" returning_customers_summary['monetary_value_cal'])"
]
},
{
"cell_type": "code",
"execution_count": 79,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"<lifetimes.GammaGammaFitter: fitted with 24766 subjects, p: 2.06, q: 0.11, v: 2.05>"
]
},
"execution_count": 79,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"ggf = GammaGammaFitter(penalizer_coef = 0.01)\n",
"ggf.fit(returning_customers_summary['frequency_cal'],\n",
" returning_customers_summary['monetary_value_cal'])"
]
},
{
"cell_type": "code",
"execution_count": 80,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"-94655.75"
]
},
"execution_count": 80,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"summary_cal_holdout_test_good.monetary_value_cal.min()"
]
},
{
"cell_type": "code",
"execution_count": 81,
"metadata": {},
"outputs": [],
"source": [
"gamma_test = summary_cal_holdout_test_good.query('frequency_cal >0 and monetary_value_cal > -50000')\n",
"gamma_test['monetary_value_cal'] = np.where(gamma_test['monetary_value_cal'] <= 0, 1, \n",
" gamma_test['monetary_value_cal'])"
]
},
{
"cell_type": "code",
"execution_count": 82,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"135865.26265577716 195861.9143043536\n",
"41014.7602962963 49535.31909758539\n",
"-59996.651648576444 -4244.689558231848\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": [
"gamma_test['monetary_pred'] = ggf.conditional_expected_average_profit(\n",
" gamma_test['frequency_cal'],\n",
" gamma_test['monetary_value_cal']\n",
")\n",
"\n",
"gamma_test['error'] = gamma_test['monetary_value_holdout'] - gamma_test['monetary_pred'] \n",
"print(gamma_test['monetary_value_holdout'].mean(), gamma_test['monetary_pred'].mean())\n",
"print(gamma_test['monetary_value_holdout'].median(), gamma_test['monetary_pred'].median())\n",
"print(gamma_test['error'].mean(), gamma_test['error'].median())\n",
"\n",
"_ = plt.hist(gamma_test['error'], bins = 50)"
]
},
{
"cell_type": "code",
"execution_count": 83,
"metadata": {},
"outputs": [],
"source": [
"# ggf.save_model('gammagamma_premium.p')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# 4. Putting It All Together"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### all numbers are based on the holdout (test) dataset"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Predict Survival Rate and Hazard Rate"
]
},
{
"cell_type": "code",
"execution_count": 84,
"metadata": {},
"outputs": [],
"source": [
"uids = summary_cal_holdout_test['user_id'].unique().tolist()"
]
},
{
"cell_type": "code",
"execution_count": 85,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(4429, 441)"
]
},
"execution_count": 85,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"test_df = test_df[test_df['set_password_timestamp'].dt.date.astype(str) >= '2018-07-01']\n",
"test_df = test_df[test_df['user_id'].isin(uids)].reset_index(drop=True)\n",
"test_df.shape"
]
},
{
"cell_type": "code",
"execution_count": 244,
"metadata": {},
"outputs": [],
"source": [
"max_duration = 16"
]
},
{
"cell_type": "code",
"execution_count": 245,
"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>user_id</th>\n",
" <th>m</th>\n",
" <th>s_rate</th>\n",
" <th>h_rate</th>\n",
" <th>s_rate_before</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>48314</th>\n",
" <td>1856381</td>\n",
" <td>1.0</td>\n",
" <td>0.998082</td>\n",
" <td>0.001918</td>\n",
" <td>1.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48315</th>\n",
" <td>1856381</td>\n",
" <td>2.0</td>\n",
" <td>0.993616</td>\n",
" <td>0.004475</td>\n",
" <td>0.998082</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48316</th>\n",
" <td>1856381</td>\n",
" <td>3.0</td>\n",
" <td>0.987637</td>\n",
" <td>0.006017</td>\n",
" <td>0.993616</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48317</th>\n",
" <td>1856381</td>\n",
" <td>4.0</td>\n",
" <td>0.980897</td>\n",
" <td>0.006824</td>\n",
" <td>0.987637</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48318</th>\n",
" <td>1856381</td>\n",
" <td>5.0</td>\n",
" <td>0.973887</td>\n",
" <td>0.007147</td>\n",
" <td>0.980897</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48319</th>\n",
" <td>1856381</td>\n",
" <td>6.0</td>\n",
" <td>0.966887</td>\n",
" <td>0.007188</td>\n",
" <td>0.973887</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48320</th>\n",
" <td>1856381</td>\n",
" <td>7.0</td>\n",
" <td>0.959949</td>\n",
" <td>0.007176</td>\n",
" <td>0.966887</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48321</th>\n",
" <td>1856381</td>\n",
" <td>8.0</td>\n",
" <td>0.953057</td>\n",
" <td>0.007179</td>\n",
" <td>0.959949</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48322</th>\n",
" <td>1856381</td>\n",
" <td>9.0</td>\n",
" <td>0.946186</td>\n",
" <td>0.007209</td>\n",
" <td>0.953057</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48323</th>\n",
" <td>1856381</td>\n",
" <td>10.0</td>\n",
" <td>0.939309</td>\n",
" <td>0.007269</td>\n",
" <td>0.946186</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48324</th>\n",
" <td>1856381</td>\n",
" <td>11.0</td>\n",
" <td>0.932399</td>\n",
" <td>0.007356</td>\n",
" <td>0.939309</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48325</th>\n",
" <td>1856381</td>\n",
" <td>12.0</td>\n",
" <td>0.925440</td>\n",
" <td>0.007464</td>\n",
" <td>0.932399</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48326</th>\n",
" <td>1856381</td>\n",
" <td>13.0</td>\n",
" <td>0.918435</td>\n",
" <td>0.007569</td>\n",
" <td>0.925440</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48327</th>\n",
" <td>1856381</td>\n",
" <td>14.0</td>\n",
" <td>0.911393</td>\n",
" <td>0.007668</td>\n",
" <td>0.918435</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48328</th>\n",
" <td>1856381</td>\n",
" <td>15.0</td>\n",
" <td>0.904320</td>\n",
" <td>0.007760</td>\n",
" <td>0.911393</td>\n",
" </tr>\n",
" <tr>\n",
" <th>48329</th>\n",
" <td>1856381</td>\n",
" <td>16.0</td>\n",
" <td>0.897224</td>\n",
" <td>0.007848</td>\n",
" <td>0.904320</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3587</th>\n",
" <td>2523994</td>\n",
" <td>1.0</td>\n",
" <td>0.998306</td>\n",
" <td>0.001694</td>\n",
" <td>1.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3588</th>\n",
" <td>2523994</td>\n",
" <td>2.0</td>\n",
" <td>0.994358</td>\n",
" <td>0.003954</td>\n",
" <td>0.998306</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3589</th>\n",
" <td>2523994</td>\n",
" <td>3.0</td>\n",
" <td>0.989071</td>\n",
" <td>0.005317</td>\n",
" <td>0.994358</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3590</th>\n",
" <td>2523994</td>\n",
" <td>4.0</td>\n",
" <td>0.983107</td>\n",
" <td>0.006030</td>\n",
" <td>0.989071</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" user_id m s_rate h_rate s_rate_before\n",
"48314 1856381 1.0 0.998082 0.001918 1.000000\n",
"48315 1856381 2.0 0.993616 0.004475 0.998082\n",
"48316 1856381 3.0 0.987637 0.006017 0.993616\n",
"48317 1856381 4.0 0.980897 0.006824 0.987637\n",
"48318 1856381 5.0 0.973887 0.007147 0.980897\n",
"48319 1856381 6.0 0.966887 0.007188 0.973887\n",
"48320 1856381 7.0 0.959949 0.007176 0.966887\n",
"48321 1856381 8.0 0.953057 0.007179 0.959949\n",
"48322 1856381 9.0 0.946186 0.007209 0.953057\n",
"48323 1856381 10.0 0.939309 0.007269 0.946186\n",
"48324 1856381 11.0 0.932399 0.007356 0.939309\n",
"48325 1856381 12.0 0.925440 0.007464 0.932399\n",
"48326 1856381 13.0 0.918435 0.007569 0.925440\n",
"48327 1856381 14.0 0.911393 0.007668 0.918435\n",
"48328 1856381 15.0 0.904320 0.007760 0.911393\n",
"48329 1856381 16.0 0.897224 0.007848 0.904320\n",
"3587 2523994 1.0 0.998306 0.001694 1.000000\n",
"3588 2523994 2.0 0.994358 0.003954 0.998306\n",
"3589 2523994 3.0 0.989071 0.005317 0.994358\n",
"3590 2523994 4.0 0.983107 0.006030 0.989071"
]
},
"execution_count": 245,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"surv_df = cph_spline.predict_survival_function(test_df[final_feats], [i for i in range(1,max_duration+1,1)]).T\n",
"\n",
"surv_df[-1.] = 1.0\n",
"surv_df['user_id'] = test_df['user_id']\n",
"\n",
"surv_df = surv_df.set_index('user_id')\n",
"\n",
"surv_df = surv_df.T.unstack().to_frame().reset_index().sort_values(by=['user_id','level_1'])\n",
"surv_df.columns = ['user_id', 'm', 's_rate']\n",
"surv_df['h_rate'] = np.abs(surv_df.groupby('user_id')['s_rate'].pct_change())\n",
"\n",
"surv_df['s_rate_before'] = surv_df['s_rate'].shift(1)\n",
"surv_df = surv_df.query(f'm >=0 and m <={max_duration}')\n",
"surv_df.head(20)"
]
},
{
"cell_type": "code",
"execution_count": 246,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"4429"
]
},
"execution_count": 246,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"surv_df['user_id'].nunique()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Predict the Frequency at each month"
]
},
{
"cell_type": "code",
"execution_count": 247,
"metadata": {},
"outputs": [],
"source": [
"for i in range(1,max_duration+1):\n",
" summary_cal_holdout_test[i] = mbgnbd.conditional_expected_number_of_purchases_up_to_time(i * 30, \n",
" summary_cal_holdout_test['frequency_cal'], \n",
" summary_cal_holdout_test['recency_cal'], \n",
" summary_cal_holdout_test['T_cal']).to_frame()"
]
},
{
"cell_type": "code",
"execution_count": 248,
"metadata": {},
"outputs": [],
"source": [
"freq_df = summary_cal_holdout_test.set_index('user_id').copy()"
]
},
{
"cell_type": "code",
"execution_count": 249,
"metadata": {},
"outputs": [],
"source": [
"freq_df = freq_df[list(range(1,max_duration+1))].copy().unstack().reset_index()\n",
"freq_df.columns = ['m','user_id','pred_freq_holdout']"
]
},
{
"cell_type": "code",
"execution_count": 250,
"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>m</th>\n",
" <th>user_id</th>\n",
" <th>pred_freq_holdout</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1</td>\n",
" <td>1856381</td>\n",
" <td>0.432692</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1</td>\n",
" <td>2523994</td>\n",
" <td>1.683893</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>1</td>\n",
" <td>2532142</td>\n",
" <td>3.057480</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>1</td>\n",
" <td>2605277</td>\n",
" <td>1.471698</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>1</td>\n",
" <td>2685026</td>\n",
" <td>0.647587</td>\n",
" </tr>\n",
" <tr>\n",
" <th>...</th>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>70859</th>\n",
" <td>16</td>\n",
" <td>5789426</td>\n",
" <td>18.002362</td>\n",
" </tr>\n",
" <tr>\n",
" <th>70860</th>\n",
" <td>16</td>\n",
" <td>5790867</td>\n",
" <td>37.008575</td>\n",
" </tr>\n",
" <tr>\n",
" <th>70861</th>\n",
" <td>16</td>\n",
" <td>5794032</td>\n",
" <td>73.524234</td>\n",
" </tr>\n",
" <tr>\n",
" <th>70862</th>\n",
" <td>16</td>\n",
" <td>5794473</td>\n",
" <td>60.601858</td>\n",
" </tr>\n",
" <tr>\n",
" <th>70863</th>\n",
" <td>16</td>\n",
" <td>5796598</td>\n",
" <td>7.164229</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>70864 rows × 3 columns</p>\n",
"</div>"
],
"text/plain": [
" m user_id pred_freq_holdout\n",
"0 1 1856381 0.432692\n",
"1 1 2523994 1.683893\n",
"2 1 2532142 3.057480\n",
"3 1 2605277 1.471698\n",
"4 1 2685026 0.647587\n",
"... .. ... ...\n",
"70859 16 5789426 18.002362\n",
"70860 16 5790867 37.008575\n",
"70861 16 5794032 73.524234\n",
"70862 16 5794473 60.601858\n",
"70863 16 5796598 7.164229\n",
"\n",
"[70864 rows x 3 columns]"
]
},
"execution_count": 250,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"freq_df"
]
},
{
"cell_type": "code",
"execution_count": 251,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"4429"
]
},
"execution_count": 251,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"freq_df['user_id'].nunique()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Predict the Monetary Value"
]
},
{
"cell_type": "code",
"execution_count": 252,
"metadata": {},
"outputs": [],
"source": [
"monetary_df = summary_cal_holdout_test.copy()"
]
},
{
"cell_type": "code",
"execution_count": 253,
"metadata": {},
"outputs": [],
"source": [
"monetary_df['pred_avg_amount_holdout'] = ggf.conditional_expected_average_profit(\n",
" summary_cal_holdout_test['frequency_cal'],\n",
" summary_cal_holdout_test['monetary_value_cal']\n",
")\n",
"\n",
"monetary_df = monetary_df[['user_id', 'pred_avg_amount_holdout']].copy()"
]
},
{
"cell_type": "code",
"execution_count": 254,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"4429"
]
},
"execution_count": 254,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"monetary_df['user_id'].nunique()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Predicted LTV\n",
"- combine all 3 models"
]
},
{
"cell_type": "code",
"execution_count": 255,
"metadata": {},
"outputs": [],
"source": [
"test_ltv_df = surv_df.merge(freq_df, how = 'left', on = ['user_id','m'])\\\n",
" .merge(monetary_df, how = 'left', on ='user_id')"
]
},
{
"cell_type": "code",
"execution_count": 213,
"metadata": {},
"outputs": [],
"source": [
"# adding additional information (actual ltv etc)\n",
"test_ltv_df = test_ltv_df.merge(main_ltv_df, how = 'left', on ='user_id')"
]
},
{
"cell_type": "code",
"execution_count": 215,
"metadata": {},
"outputs": [],
"source": [
"test_ltv_df['pred_freq_holdout_inc'] = test_ltv_df.groupby('user_id')['pred_freq_holdout'].diff()\n",
"test_ltv_df['pred_freq_holdout_inc'] = np.where(test_ltv_df['pred_freq_holdout_inc'].isna(), \n",
" test_ltv_df['pred_freq_holdout'], test_ltv_df['pred_freq_holdout_inc'])"
]
},
{
"cell_type": "code",
"execution_count": 216,
"metadata": {},
"outputs": [],
"source": [
"test_ltv_df['pred_rev_holdout'] = (test_ltv_df['pred_freq_holdout_inc'] * test_ltv_df['pred_avg_amount_holdout'] * test_ltv_df['s_rate']).astype(int)\n",
"test_ltv_df['pred_npl_holdout'] = (test_ltv_df['snapshot_os_principal'] * test_ltv_df['h_rate'] * test_ltv_df['s_rate_before'] * 0.9).astype(int)\n",
"test_ltv_df['pred_ltv_holdout'] = test_ltv_df['pred_rev_holdout'] - test_ltv_df['pred_npl_holdout']"
]
},
{
"cell_type": "code",
"execution_count": 217,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>user_id</th>\n",
" <th>m</th>\n",
" <th>s_rate</th>\n",
" <th>h_rate</th>\n",
" <th>s_rate_before</th>\n",
" <th>pred_freq_holdout</th>\n",
" <th>pred_avg_amount_holdout</th>\n",
" <th>ltv_calib</th>\n",
" <th>act_ltv_holdout</th>\n",
" <th>act_npl_holdout</th>\n",
" <th>act_rev_holdout</th>\n",
" <th>snapshot_os_principal</th>\n",
" <th>pred_freq_holdout_inc</th>\n",
" <th>pred_rev_holdout</th>\n",
" <th>pred_npl_holdout</th>\n",
" <th>pred_ltv_holdout</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1856381</td>\n",
" <td>1.0</td>\n",
" <td>0.998082</td>\n",
" <td>0.001918</td>\n",
" <td>1.000000</td>\n",
" <td>0.432692</td>\n",
" <td>1.023064e+06</td>\n",
" <td>675198.96</td>\n",
" <td>541075.77</td>\n",
" <td>0.0</td>\n",
" <td>541075.77</td>\n",
" <td>2541746.0</td>\n",
" <td>0.432692</td>\n",
" <td>441822</td>\n",
" <td>4386</td>\n",
" <td>437436</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1856381</td>\n",
" <td>2.0</td>\n",
" <td>0.993616</td>\n",
" <td>0.004475</td>\n",
" <td>0.998082</td>\n",
" <td>0.863352</td>\n",
" <td>1.023064e+06</td>\n",
" <td>675198.96</td>\n",
" <td>541075.77</td>\n",
" <td>0.0</td>\n",
" <td>541075.77</td>\n",
" <td>2541746.0</td>\n",
" <td>0.430659</td>\n",
" <td>437779</td>\n",
" <td>10217</td>\n",
" <td>427562</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>1856381</td>\n",
" <td>3.0</td>\n",
" <td>0.987637</td>\n",
" <td>0.006017</td>\n",
" <td>0.993616</td>\n",
" <td>1.292334</td>\n",
" <td>1.023064e+06</td>\n",
" <td>675198.96</td>\n",
" <td>541075.77</td>\n",
" <td>0.0</td>\n",
" <td>541075.77</td>\n",
" <td>2541746.0</td>\n",
" <td>0.428982</td>\n",
" <td>433450</td>\n",
" <td>13677</td>\n",
" <td>419773</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>1856381</td>\n",
" <td>4.0</td>\n",
" <td>0.980897</td>\n",
" <td>0.006824</td>\n",
" <td>0.987637</td>\n",
" <td>1.719890</td>\n",
" <td>1.023064e+06</td>\n",
" <td>675198.96</td>\n",
" <td>541075.77</td>\n",
" <td>0.0</td>\n",
" <td>541075.77</td>\n",
" <td>2541746.0</td>\n",
" <td>0.427556</td>\n",
" <td>429061</td>\n",
" <td>15416</td>\n",
" <td>413645</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>1856381</td>\n",
" <td>5.0</td>\n",
" <td>0.973887</td>\n",
" <td>0.007147</td>\n",
" <td>0.980897</td>\n",
" <td>2.146208</td>\n",
" <td>1.023064e+06</td>\n",
" <td>675198.96</td>\n",
" <td>541075.77</td>\n",
" <td>0.0</td>\n",
" <td>541075.77</td>\n",
" <td>2541746.0</td>\n",
" <td>0.426318</td>\n",
" <td>424761</td>\n",
" <td>16035</td>\n",
" <td>408726</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" user_id m s_rate h_rate s_rate_before pred_freq_holdout \\\n",
"0 1856381 1.0 0.998082 0.001918 1.000000 0.432692 \n",
"1 1856381 2.0 0.993616 0.004475 0.998082 0.863352 \n",
"2 1856381 3.0 0.987637 0.006017 0.993616 1.292334 \n",
"3 1856381 4.0 0.980897 0.006824 0.987637 1.719890 \n",
"4 1856381 5.0 0.973887 0.007147 0.980897 2.146208 \n",
"\n",
" pred_avg_amount_holdout ltv_calib act_ltv_holdout act_npl_holdout \\\n",
"0 1.023064e+06 675198.96 541075.77 0.0 \n",
"1 1.023064e+06 675198.96 541075.77 0.0 \n",
"2 1.023064e+06 675198.96 541075.77 0.0 \n",
"3 1.023064e+06 675198.96 541075.77 0.0 \n",
"4 1.023064e+06 675198.96 541075.77 0.0 \n",
"\n",
" act_rev_holdout snapshot_os_principal pred_freq_holdout_inc \\\n",
"0 541075.77 2541746.0 0.432692 \n",
"1 541075.77 2541746.0 0.430659 \n",
"2 541075.77 2541746.0 0.428982 \n",
"3 541075.77 2541746.0 0.427556 \n",
"4 541075.77 2541746.0 0.426318 \n",
"\n",
" pred_rev_holdout pred_npl_holdout pred_ltv_holdout \n",
"0 441822 4386 437436 \n",
"1 437779 10217 427562 \n",
"2 433450 13677 419773 \n",
"3 429061 15416 413645 \n",
"4 424761 16035 408726 "
]
},
"execution_count": 217,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"test_ltv_df.head()"
]
},
{
"cell_type": "code",
"execution_count": 218,
"metadata": {},
"outputs": [],
"source": [
"test_agg_ltv_df = test_ltv_df.groupby('user_id').agg({\n",
" 'ltv_calib': max,\n",
" 'act_ltv_holdout': max,\n",
" 'act_npl_holdout': max,\n",
" 'act_rev_holdout': max,\n",
" 'pred_rev_holdout': sum,\n",
" 'pred_npl_holdout': sum,\n",
" 'pred_ltv_holdout': sum}).reset_index()"
]
},
{
"cell_type": "code",
"execution_count": 219,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>user_id</th>\n",
" <th>ltv_calib</th>\n",
" <th>act_ltv_holdout</th>\n",
" <th>act_npl_holdout</th>\n",
" <th>act_rev_holdout</th>\n",
" <th>pred_rev_holdout</th>\n",
" <th>pred_npl_holdout</th>\n",
" <th>pred_ltv_holdout</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1856381</td>\n",
" <td>6.751990e+05</td>\n",
" <td>5.410758e+05</td>\n",
" <td>0.0</td>\n",
" <td>5.410758e+05</td>\n",
" <td>6588076</td>\n",
" <td>235099</td>\n",
" <td>6352977</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>2523994</td>\n",
" <td>9.852474e+05</td>\n",
" <td>1.738388e+06</td>\n",
" <td>0.0</td>\n",
" <td>1.738388e+06</td>\n",
" <td>1121826</td>\n",
" <td>196819</td>\n",
" <td>925007</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>2532142</td>\n",
" <td>1.094385e+06</td>\n",
" <td>2.849916e+06</td>\n",
" <td>0.0</td>\n",
" <td>2.849916e+06</td>\n",
" <td>2344182</td>\n",
" <td>182443</td>\n",
" <td>2161739</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>2605277</td>\n",
" <td>1.325461e+06</td>\n",
" <td>2.869269e+06</td>\n",
" <td>0.0</td>\n",
" <td>2.869269e+06</td>\n",
" <td>683184</td>\n",
" <td>115787</td>\n",
" <td>567397</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>2685026</td>\n",
" <td>4.759348e+06</td>\n",
" <td>3.665706e+06</td>\n",
" <td>0.0</td>\n",
" <td>3.665706e+06</td>\n",
" <td>14045413</td>\n",
" <td>1052526</td>\n",
" <td>12992887</td>\n",
" </tr>\n",
" <tr>\n",
" <th>...</th>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4424</th>\n",
" <td>5789426</td>\n",
" <td>4.320403e+04</td>\n",
" <td>1.475467e+05</td>\n",
" <td>0.0</td>\n",
" <td>1.475467e+05</td>\n",
" <td>174553</td>\n",
" <td>1480</td>\n",
" <td>173073</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4425</th>\n",
" <td>5790867</td>\n",
" <td>1.662109e+06</td>\n",
" <td>1.571261e+06</td>\n",
" <td>0.0</td>\n",
" <td>1.571261e+06</td>\n",
" <td>5989083</td>\n",
" <td>576588</td>\n",
" <td>5412495</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4426</th>\n",
" <td>5794032</td>\n",
" <td>2.486986e+05</td>\n",
" <td>1.077321e+06</td>\n",
" <td>0.0</td>\n",
" <td>1.077321e+06</td>\n",
" <td>599958</td>\n",
" <td>71459</td>\n",
" <td>528499</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4427</th>\n",
" <td>5794473</td>\n",
" <td>1.663013e+05</td>\n",
" <td>7.407890e+04</td>\n",
" <td>0.0</td>\n",
" <td>7.407890e+04</td>\n",
" <td>1009084</td>\n",
" <td>38576</td>\n",
" <td>970508</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4428</th>\n",
" <td>5796598</td>\n",
" <td>9.695000e+05</td>\n",
" <td>3.233113e+06</td>\n",
" <td>0.0</td>\n",
" <td>3.233113e+06</td>\n",
" <td>5946292</td>\n",
" <td>122416</td>\n",
" <td>5823876</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>4429 rows × 8 columns</p>\n",
"</div>"
],
"text/plain": [
" user_id ltv_calib act_ltv_holdout act_npl_holdout \\\n",
"0 1856381 6.751990e+05 5.410758e+05 0.0 \n",
"1 2523994 9.852474e+05 1.738388e+06 0.0 \n",
"2 2532142 1.094385e+06 2.849916e+06 0.0 \n",
"3 2605277 1.325461e+06 2.869269e+06 0.0 \n",
"4 2685026 4.759348e+06 3.665706e+06 0.0 \n",
"... ... ... ... ... \n",
"4424 5789426 4.320403e+04 1.475467e+05 0.0 \n",
"4425 5790867 1.662109e+06 1.571261e+06 0.0 \n",
"4426 5794032 2.486986e+05 1.077321e+06 0.0 \n",
"4427 5794473 1.663013e+05 7.407890e+04 0.0 \n",
"4428 5796598 9.695000e+05 3.233113e+06 0.0 \n",
"\n",
" act_rev_holdout pred_rev_holdout pred_npl_holdout pred_ltv_holdout \n",
"0 5.410758e+05 6588076 235099 6352977 \n",
"1 1.738388e+06 1121826 196819 925007 \n",
"2 2.849916e+06 2344182 182443 2161739 \n",
"3 2.869269e+06 683184 115787 567397 \n",
"4 3.665706e+06 14045413 1052526 12992887 \n",
"... ... ... ... ... \n",
"4424 1.475467e+05 174553 1480 173073 \n",
"4425 1.571261e+06 5989083 576588 5412495 \n",
"4426 1.077321e+06 599958 71459 528499 \n",
"4427 7.407890e+04 1009084 38576 970508 \n",
"4428 3.233113e+06 5946292 122416 5823876 \n",
"\n",
"[4429 rows x 8 columns]"
]
},
"execution_count": 219,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"test_agg_ltv_df"
]
},
{
"cell_type": "code",
"execution_count": 220,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"552618.1751571696 1356748.2467825694\n",
"948877.4339999998 1072792.0\n"
]
}
],
"source": [
"print(test_agg_ltv_df['act_ltv_holdout'].mean(), test_agg_ltv_df['pred_ltv_holdout'].mean())\n",
"print(test_agg_ltv_df['act_ltv_holdout'].median(), test_agg_ltv_df['pred_ltv_holdout'].median()) "
]
},
{
"cell_type": "code",
"execution_count": 221,
"metadata": {},
"outputs": [],
"source": [
"test_agg_ltv_df['ltv_act'] = test_agg_ltv_df['ltv_calib'] + test_agg_ltv_df['act_ltv_holdout']\n",
"test_agg_ltv_df['ltv_pred'] = test_agg_ltv_df['ltv_calib'] + test_agg_ltv_df['pred_ltv_holdout']"
]
},
{
"cell_type": "code",
"execution_count": 222,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"1710470.8833193057 2514600.954944705\n",
"1764373.3979999998 1892217.84\n"
]
}
],
"source": [
"print(test_agg_ltv_df['ltv_act'].mean(), test_agg_ltv_df['ltv_pred'].mean())\n",
"print(test_agg_ltv_df['ltv_act'].median(), test_agg_ltv_df['ltv_pred'].median()) "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Evaluation"
]
},
{
"cell_type": "code",
"execution_count": 223,
"metadata": {},
"outputs": [],
"source": [
"# convert the ltv into dollars\n",
"# divide by 14000 (dollar to rupiah exchange rate at the time)\n",
"\n",
"test_agg_ltv_df['bin_act'], _bins = pd.qcut(test_agg_ltv_df['ltv_act']/14000, 5, retbins = True)\n",
"test_agg_ltv_df['bin_pred'] = pd.cut(test_agg_ltv_df['ltv_pred']/14000, bins = _bins)"
]
},
{
"cell_type": "code",
"execution_count": 224,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([-1354.86612754, 5.77198926, 90.29589965, 162.01937746,\n",
" 256.28726919, 1593.64090251])"
]
},
"execution_count": 224,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"_bins"
]
},
{
"cell_type": "code",
"execution_count": 227,
"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>bin_pred</th>\n",
" <th>(-1354.866, 5.772]</th>\n",
" <th>(5.772, 90.296]</th>\n",
" <th>(90.296, 162.019]</th>\n",
" <th>(162.019, 256.287]</th>\n",
" <th>(256.287, 1593.641]</th>\n",
" </tr>\n",
" <tr>\n",
" <th>bin_act</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>(-1354.867, 5.772]</th>\n",
" <td>350</td>\n",
" <td>156</td>\n",
" <td>91</td>\n",
" <td>103</td>\n",
" <td>182</td>\n",
" </tr>\n",
" <tr>\n",
" <th>(5.772, 90.296]</th>\n",
" <td>76</td>\n",
" <td>512</td>\n",
" <td>170</td>\n",
" <td>81</td>\n",
" <td>47</td>\n",
" </tr>\n",
" <tr>\n",
" <th>(90.296, 162.019]</th>\n",
" <td>40</td>\n",
" <td>298</td>\n",
" <td>251</td>\n",
" <td>176</td>\n",
" <td>120</td>\n",
" </tr>\n",
" <tr>\n",
" <th>(162.019, 256.287]</th>\n",
" <td>22</td>\n",
" <td>149</td>\n",
" <td>201</td>\n",
" <td>261</td>\n",
" <td>253</td>\n",
" </tr>\n",
" <tr>\n",
" <th>(256.287, 1593.641]</th>\n",
" <td>23</td>\n",
" <td>83</td>\n",
" <td>116</td>\n",
" <td>152</td>\n",
" <td>504</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
"bin_pred (-1354.866, 5.772] (5.772, 90.296] (90.296, 162.019] \\\n",
"bin_act \n",
"(-1354.867, 5.772] 350 156 91 \n",
"(5.772, 90.296] 76 512 170 \n",
"(90.296, 162.019] 40 298 251 \n",
"(162.019, 256.287] 22 149 201 \n",
"(256.287, 1593.641] 23 83 116 \n",
"\n",
"bin_pred (162.019, 256.287] (256.287, 1593.641] \n",
"bin_act \n",
"(-1354.867, 5.772] 103 182 \n",
"(5.772, 90.296] 81 47 \n",
"(90.296, 162.019] 176 120 \n",
"(162.019, 256.287] 261 253 \n",
"(256.287, 1593.641] 152 504 "
]
},
"execution_count": 227,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"test_agg_ltv_df.groupby('bin_act')['bin_pred'].value_counts().unstack()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.10"
}
},
"nbformat": 4,
"nbformat_minor": 2
}