Feature Scaling Visualizer¶

View Fullscreen

Description¶

An interactive visualization comparing min-max scaling and z-score standardization across different data distributions.

Learning Objectives¶

• Understand how min-max scaling transforms data to the [0, 1] range
• Compare z-score standardization that creates mean=0, std=1 distributions
• Observe how outliers affect each scaling method differently
• Recognize when to use each scaling technique

How to Use¶

1. Select Distribution: Choose from Normal, Skewed, With Outliers, or Bimodal distributions
2. Adjust Parameters: Use sliders to change sample size, mean, and standard deviation
3. Add Outliers: Click the button to add outliers and observe their impact
4. Compare Results: Examine histograms, box plots, and statistics across all three panels

Key Concepts¶

Min-Max Scaling¶

• Transforms to [0, 1] range
• Formula: x' = (x - min) / (max - min)
• Highly sensitive to outliers
• Preserves distribution shape

Z-Score Standardization¶

• Transforms to mean=0, std=1
• Formula: x' = (x - μ) / σ
• Less sensitive to outliers
• Assumes approximately Gaussian distribution

Interactive Features¶

• Distribution Types: Compare scaling effects on different data shapes
• Live Statistics: View mean, std, min, max for each transformation
• Box Plots: Visualize quartiles and outliers
• Histograms: See the distribution shape before and after scaling

Related Concepts¶

• Normalization
• Standardization
• Data Preprocessing