Interactive Statistics Studio

Rivu Basu PCA lab

Learn PCA by building it with your own hands. Draw or generate data, center and standardize it, inspect the covariance matrix, watch eigenvectors emerge, and then step into a rotatable 3D scene where PCA finds the best low-dimensional summary of a cloud in space.

Points 0
Mode 2D
Status Collect Data
Step 1 / 6
0.35
35 deg
24 deg

Data Space

Click or drag to spray 2D points
Raw observations Mean

Analysis Space

Standardized coordinates and PCA scores
PC1 direction / score PC2 direction / score

3D Explorer

Use sliders or drag to rotate the scene
PC1 axis PC2 axis PC3 axis
Current Lesson

Step 1. Meet the raw data cloud

Start by making a dataset that has some shape. PCA is only interesting when there is structure to explain.

Live Metrics

Mean vector [0.000, 0.000]
Standard deviations [1.000, 1.000]
Covariance off-diagonal 0.000
Projection state Hidden

Explained Variance

PC1: 0.0%
PC2: 0.0%
PC3: 0.0%

Numerical Readout

Add points or load a preset to begin the tutorial.

PCA Walkthrough