midPCA
What is the geometric interpretation of PCA?
Updated May 17, 2026
Short answer
PCA rotates coordinate axes to align with directions of maximum variance.
Deep explanation
Geometrically, PCA finds orthogonal axes that best fit data spread. First component captures maximum variance direction, second is orthogonal and captures next highest variance, and so on. It is essentially a rotation and projection of data space.
Real-world example
Visualizing high-dimensional customer data in 2D.
Common mistakes
- Thinking PCA distorts data instead of rotating it.
Follow-up questions
- Does PCA preserve distances?
- Is PCA linear transformation?