pro3

preprocess_pro3.py

import numpy as np
import pandas as pd
import sklearn as sk
from sklearn.preprocessing import LabelEncoder

data = pd.read_csv('raw_data.csv') #I had renamed the downloaded file as raw_data.csv, please check out the actual file name in your case
data.drop(['identifierHash','type','country','language','civilityTitle','civilityGenderId','seniorityAsMonths','seniorityAsYears','countryCode','hasAndroidApp','hasIosApp'],axis=1,inplace=True)
def pure_round(num):
    integer = int(num)
    fraction = num-float(integer)
    if fraction>=0.5:
        integer+=1
    return integer
  
  
for i in data.index:
    case_no = data.loc[i,'productsSold']
    pass_no = pure_round((case_no * data.loc[i,'productsPassRate']) / 100)
    fail_no = case_no - pass_no
    data.loc[i,'productsPassed'] = pass_no
    data.loc[i,'productsFailed'] = fail_no
    if case_no == 0:
        data.drop(i, axis=0,inplace=True)
 
data.drop(['productsPassRate','productsSold'], axis=1,inplace=True)


Encodedict={}
for i in ['gender','hasAnyApp','hasProfilePicture']:
    key='_{}'.format(i)
    le=LabelEncoder()
    data[key]=le.fit_transform(list(data[i]))
    Encodedict[key]=le.classes_                          
                              
    
data.drop(['gender','hasAnyApp','hasProfilePicture'],axis=1,inplace=True)
indices = data3.index
dfdict = {}
for j in indices:
    x = data3.loc[j,'productsPassed']
    y = data3.loc[j,'productsFailed']
    if x != 0:
        data3.loc[j,'Fraud'] = 0
        df = pd.DataFrame(data3.loc[j,:]).transpose()
        ldf = pd.concat([df]*int(x), ignore_index=True)
    
    if y != 0:
        data3.loc[j,'Fraud'] = 1
        df2 = pd.DataFrame(data3.loc[j,:]).transpose()
        ldf2 = pd.concat([df2]*int(y), ignore_index=True)
    
    if x!=0 and y!=0:
        dfdict[i] = pd.concat([ldf, ldf2],ignore_index=True)
    elif x!=0:
        dfdict[i] = ldf
    else:
        dfdict[i] = ldf2
        

whole_df = pd.concat(dfdict.values(),ignore_index=True)
whole_df.drop(['productsPassed','productsFailed'],axis=1,inplace=True)
whole_df.to_csv('modified_data.csv')