why the calculation of eigenvectors and eigenvalues when performing PCA is so effective?

Question

The core of Principal Componenet Analysis (PCA) lies at calculating eigenvalues and eigenvectors from the variance-covariance matrix corresponding to some dataset (for example,