pandas_on_spark_monthly_loan_stats.py

# group by month using the loan created date and calculate some stats off the grouped data
def group_functions(x):
    d = {}
    
    d['loan_app_count'] = x['loan_amount'].count()
    d['loan_funded_count'] = x['funded_at'].count()
    d['conversion_rate'] = max(x['funded_at'].count() /x['created_at'].count(), 0)
    d['time_to_conversion_avg'] = x['time_to_conversion'].mean()
    
    return pd.Series(d, index=['loan_app_count', 'loan_funded_count', 
                                'conversion_rate', 'time_to_conversion_avg'])
                                
monthly = loans_df.created_at.dt.strftime('%Y-%m')
monthly_loan_stats = loans_df.groupby(monthly).apply(group_functions)
monthly_loan_stats.index.rename('created_at_month', inplace=True)