PCA Visualizer

Overview

The PCA Visualizer demystifies Principal Component Analysis through interactive visualizations. Transform high-dimensional data into 2D or 3D principal component space, explore scree plots showing explained variance, examine loading vectors to understand which original features matter most, and create biplots showing both observations and variables. This tool makes dimensionality reduction intuitive and helps you understand which components capture the most information in your data.

Open in new tab

Tips

  • Start with 3D data to see how PCA finds the best 2D projection
  • The scree plot shows explained variance by component - look for the “elbow” where variance drops off
  • PC1 (first component) always captures the most variance, PC2 the second most, and so on
  • Loading vectors show how original features contribute to each principal component
  • Use cumulative explained variance to decide how many components to keep (often aim for 80-95%)
  • Generate new data to see how PCA adapts to different correlation structures
  • Notice that PCA finds orthogonal (perpendicular) directions of maximum variance
  • The biplot combines data points and variable arrows - longer arrows mean more important variables
  • PCA is sensitive to scale - features with larger ranges dominate unless data is standardized
  • Components are linear combinations of original features, making them interpretable