PCA (principal component analysis)

Also written: PCA, principal components, principal component analysis

Finding new axes for data, ordered by how much the data spreads along each, so keeping the first two gives a 2D map that loses little. It standardizes, builds the covariance matrix and takes eigenvectors.