How to use the PCA Calculator
- 1. Paste numeric dataOne row per item and one column per feature; a header row is optional.
- 2. Choose optionsStandardise features if their units differ, and pick how many components to show.
- 3. Run PCARead eigenvalues, explained variance, loadings and scores, and download the scores.
What PCA does
Principal component analysis finds new axes, the principal components, that point along the directions of greatest variation in your data. PC1 captures the most variance, PC2 the most of what's left (at right angles to PC1), and so on. The tool centres each column (and, if you choose, divides by its standard deviation), builds the covariance matrix and finds its eigenvalues and eigenvectors with the Jacobi method. Eigenvalues show how much variance each component holds; eigenvectors (loadings) show how each original feature contributes.
Worked example
The sample has marks for 10 students in Maths, Physics, Chemistry, English and History. Standardised, PC1 explains 93.80% of the variance (eigenvalue 4.6901) and PC2 a further 4.76%, so two components keep 98.56%. PC1's loadings are positive for the science subjects and negative for English and History. It's a single "science versus humanities" axis, which is why one component is enough to pass 90% here.
Good practice
- Standardise when columns use different units (rupees, years, marks); otherwise the largest-scale column dominates.
- A component's sign is arbitrary. We make each one's largest loading positive, as scikit-learn does.
- Eigenvalues match scikit-learn's
explained_variance_, checked on 20 random datasets. - PCA finds straight-line structure only, and it isn't feature selection: every component mixes all features.
Frequently asked questions
How many components should I keep?
Common rules are enough components to reach 80–95% of the variance, or the point where the eigenvalues level off (the elbow). It depends on what you'll use them for.
How much data can I paste?
Up to 5,000 rows and 30 columns; it runs in your browser.
Is the PCA calculator free?
Yes. You can open the PCA calculator and enter everything without signing up. When you run PCA for the first time, we ask once for your name and WhatsApp number; after that it, and every other Uncodemy tool, opens straight away.