um_kod_zm_kat_3.py

# Import libraries
import numpy as np
import pandas as pd

# Display all columns
pd.set_option('display.max_columns', None)

# Import Houseprice data from GitHub
data = pd.read_csv('https://github.com/jurand71/datasets/raw/master/HouseSalePriceCompetition/houseprice.csv')

# Three variables were chosen from categorical variables for OneHotEncoder
usecols = ['Neighborhood','Exterior1st','Exterior2nd']
data = data[usecols]

# How many categories are in selected variables
for col in usecols:
    print(col,': ',len(data[col].unique()))

# Find top categories and encode with one-hot encoding method
def find_top_categories(df, variable, count=10):
    return [
        x for x in df[variable].value_counts().sort_values(ascending=False).head(count).index
    ]

def onehot_encode(df, variable, top_categories):
    for label in top_categories:
        df[variable + '_' + label] = np.where(df[variable] == label,1,0)

for var in usecols:
    top_categories = find_top_categories(data, var)
    onehot_encode(data, var, top_categories)

data