This document discusses using principal component analysis (PCA) to model the signature of manufacturing processes based on machined profile measurements. PCA is presented as a statistical technique that can identify patterns in multivariate data without requiring a parametric model. The document outlines how PCA works and applies it to real roundness measurement data from turned cylindrical parts to investigate using PCA for process signature identification. PCA is shown to effectively describe the variability observed across profiles in a way that summarizes most information with a small number of principal components.