PCA dimensionality reduction.
Find the directions of maximum variance and see a 3D dataset projected into two dimensions.
Your experiment starts here
Ready to explore?
Runs in your browser.
Quick start
Get hands-on.
Adjust the point count, cluster count, and spread. Generate data, run PCA, and compare the 3D view with its 2D projection.
What to look for
- Mean-centering and covariance-matrix computation.
- Power iteration with deflation to estimate principal components.
- Explained-variance ratios and projection onto the leading eigenvectors.