Scaling Numerical Features Using RobustScaler in Pandas

Last update on May 06 2025 13:19:25 (UTC/GMT +8 hours)

15. Scaling Numerical Features Using Scikit-learn's RobustScaler

Write a Pandas program to scale numerical features using Scikit-learn's RobustScaler.

This exercise shows how to scale numerical features using Scikit-learn's RobustScaler to reduce the effect of outliers.

Sample Solution :

Code :

import pandas as pd
from sklearn.preprocessing import RobustScaler

# Load the dataset
df = pd.read_csv('data.csv')

# Initialize the RobustScaler
scaler = RobustScaler()

# Apply RobustScaler to the 'Age' and 'Salary' columns
df[['Age', 'Salary']] = scaler.fit_transform(df[['Age', 'Salary']])

# Output the scaled dataset
print(df)

Output:

   ID      Name  Age  Gender    Salary  Target
0   1      Sara -0.8  Female -0.666667       0
1   2    Ophrah  0.2    Male  0.000000       1
2   3    Torben -1.4    Male  0.666667       0
3   4  Masaharu  1.2    Male  1.333333       1
4   5      Kaya  NaN  Female -0.333333       0
5   6   Abaddon  0.0    Male       NaN       1

Explanation:

Loaded the dataset using Pandas.
Initialized RobustScaler to scale features while reducing the influence of outliers.
Applied RobustScaler to the 'Age' and 'Salary' columns.
Displayed the scaled dataset.

For more Practice: Solve these Related Problems:

Write a Pandas program to scale numerical features using RobustScaler and compare the results with standard scaling.
Write a Pandas program to apply RobustScaler to a DataFrame and verify that the scaled features are less affected by outliers.
Write a Pandas program to use RobustScaler on selected columns and plot the distribution before and after scaling.
Write a Pandas program to combine RobustScaler with other preprocessing steps and output a transformed dataset.

Go to:

Previous: Applying Polynomial Features for Feature Expansion.
Next: Saving the Processed Dataset to a CSV File.

Python-Pandas Code Editor:

Have another way to solve this solution? Contribute your code (and comments) through Disqus.