Python Script for Replacing Missing Data in a Machine Learning Algorithm

The Python script below leverages the Imputer library and transforms a data set so that it takes the column average instead of allowing NaN values.

import numpy as np # Mathematical tools
import matplotlib.pyplot as plt # Plotting/visualization tools
import pandas as pd # Data set management tool
raw_data = pd.read_csv('raw_data.csv')
features = raw_data.iloc[:, :-1].values
dependent_variable = raw_data.iloc[:, 4].values
# Fixing missing data
from sklearn.preprocessing import Imputer
imputer = Imputer(missing_values = "NaN", strategy = 'mean', axis = 0) # Options for strategy include: mean, median, and most_frequent
imputer = imputer.fit(features[:, 0:2])
features[:, 0:2] = imputer.transform(features[:, 0:2])