按行 pivot 表占总数的百分比

Question

Im am using the following function:

def pivot_count(df, rows, columns, calc_field):
    df_pivot = df.pivot_table(values=calc_field, 
                              index=rows, 
                              columns=columns, 
                              aggfunc=np.size
                             ).dropna(axis=0, how='all')
    return df_pivot

To create a pivot table with the count of each 'calc_field' to get something like this:

ProfitCluster   Affluent Customer   High Net Worth  Mass Customer
HighMargin             302                324           568
HighRevenue            301                323           645
LessProfitable         246                246           529 

How can i modify the function to get something like this:

ProfitCluster   Affluent Customer   High Net Worth  Mass Customer
HighMargin             0.25               0.27          0.47
HighRevenue            0.23               0.25          0.50
LessProfitable         0.24               0.24          0.51 

Thanks, SOA community.

Answer 1

Change to crossstab

def pivot_count(df, rows, columns, calc_field):
    df_pivot = pd.crossstab(values=df[calc_field], 
                              index=df[rows], 
                              columns=df[columns], 
                              normalize='index',
                              aggfunc=np.size
                             )
    return df_pivot