Principal components analysis is a technique to find patterns of correlation between many possible variables or subsets of data, then reduce them to a smaller, more-manageable number of components.

Maybe only 2 components out of a dozen variables (usually set at 5) explain 95% of the variance. Use a computer program like Minitab to crunch it if you don’t want to do the linear algebra and matrix work.

In the end you’re comparing Eigenvalues to each other to decide which variables warrant greater attention.