Shows a 2D correlated point cloud, computes its covariance matrix, and animates how PCA finds orthogonal eigenvector axes (v1, v2). Points then project onto v1 to illustrate dimensionality reduction, while eigenvalues (λ1, λ2) and a bar chart show explained variance ordering.
Generates a deterministic elongated dataset, computes 2×2 covariance and analytic eigenvalues/eigenvector, then animates (1) axis rotation toward v1, (2) projection of points onto v1, and (3) k=1 reconstruction line plus explained-variance bars. Uses green-on-black, snapped coordinates for a blocky aesthetic, and time-based easing with a ~4.2s loop.