get_agg_by_cols.py

# function to calculate feature engineering by aggregating integer columns
def get_agg_by_cols(data, col, agg_list, agg_cols):
    """
    data : dataframe to calculate FE on
    col : list of columns to aggregate by (the object datatype)
    agg_list : a list of statistical measure e.g mean, median
    agg_cols : a list of numerical datatype columns
    """
    for cols in col:
        for i in agg_list:
            for j in agg_cols:
                data[cols+'_'+j+'_'+i] = data.groupby([cols])[j].transform(i)
    return data