monthsSinceRegDate.py

import pandas as pd
import numpy as np
from datetime import date
from functions import aux_functions

curr_date = pd.to_datetime(date.today())

hardbounce_2['monthsSinceRegDate'] = (curr_date - hardbounce_2['regDate_n']) / np.timedelta64(1, 'M')

# There are rows with monthsSinceRegDate missing
aux_functions.percMissing(hardbounce_2)

# There are rows with months less than 1
hardbounce_2['monthsSinceRegDate'][hardbounce_2['monthsSinceRegDate'].notnull()].astype(int).describe()

# Fill with 1 if the value < 0
hardbounce_2.loc[hardbounce_2.monthsSinceRegDate < 0, 'monthsSinceRegDate'] = 1

# Fill with median whem missing
print('Median: ' + str(hardbounce_2['monthsSinceRegDate'].median()))
hardbounce_2['monthsSinceRegDate'] = hardbounce_2['monthsSinceRegDate'].fillna(hardbounce_2['monthsSinceRegDate'].median())

# Convert to int
hardbounce_2['monthsSinceRegDate'] = hardbounce_2['monthsSinceRegDate'].astype(int)

# Describe monthsSinceRegDate
hardbounce_2['monthsSinceRegDate'].describe()