preprocessing.py

import pandas as pd
from sklearn.preprocessing import MinMaxScaler

def preprocess(df, option):
    """
    This function is to cover all the preprocessing steps on the churn dataframe. It involves selecting important features, encoding categorical data, handling missing values,feature scaling and splitting the data
    """
    #Defining the map function
    def binary_map(feature):
        return feature.map({'Yes':1, 'No':0})

    # Encode binary categorical features
    binary_list = ['SeniorCitizen','Dependents', 'PhoneService', 'PaperlessBilling']
    df[binary_list] = df[binary_list].apply(binary_map)

    
    #Drop values based on operational options
    if (option == "Online"):
        columns = ['SeniorCitizen', 'Dependents', 'tenure', 'PhoneService', 'PaperlessBilling', 'MonthlyCharges', 'TotalCharges', 'MultipleLines_No_phone_service', 'MultipleLines_Yes', 'InternetService_Fiber_optic', 'InternetService_No', 'OnlineSecurity_No_internet_service', 'OnlineSecurity_Yes', 'OnlineBackup_No_internet_service', 'TechSupport_No_internet_service', 'TechSupport_Yes', 'StreamingTV_No_internet_service', 'StreamingTV_Yes', 'StreamingMovies_No_internet_service', 'StreamingMovies_Yes', 'Contract_One_year', 'Contract_Two_year', 'PaymentMethod_Electronic_check']
        #Encoding the other categorical categoric features with more than two categories
        df = pd.get_dummies(df).reindex(columns=columns, fill_value=0)
    elif (option == "Batch"):
        pass
        df = df[['SeniorCitizen','Dependents','tenure','PhoneService','MultipleLines','InternetService','OnlineSecurity',
                'OnlineBackup','TechSupport','StreamingTV','StreamingMovies','Contract','PaperlessBilling','PaymentMethod',
                'MonthlyCharges','TotalCharges']]
        columns = ['SeniorCitizen', 'Dependents', 'tenure', 'PhoneService', 'PaperlessBilling', 'MonthlyCharges', 'TotalCharges', 'MultipleLines_No_phone_service', 'MultipleLines_Yes', 'InternetService_Fiber_optic', 'InternetService_No', 'OnlineSecurity_No_internet_service', 'OnlineSecurity_Yes', 'OnlineBackup_No_internet_service', 'TechSupport_No_internet_service', 'TechSupport_Yes', 'StreamingTV_No_internet_service', 'StreamingTV_Yes', 'StreamingMovies_No_internet_service', 'StreamingMovies_Yes', 'Contract_One_year', 'Contract_Two_year', 'PaymentMethod_Electronic_check']
        #Encoding the other categorical categoric features with more than two categories
        df = pd.get_dummies(df).reindex(columns=columns, fill_value=0)
    else:
        print("Incorrect operational options")


    #feature scaling
    sc = MinMaxScaler()
    df['tenure'] = sc.fit_transform(df[['tenure']])
    df['MonthlyCharges'] = sc.fit_transform(df[['MonthlyCharges']])
    df['TotalCharges'] = sc.fit_transform(df[['TotalCharges']])
    return df