Abstract
An algorithm is developed to help identify the sources of dimensional variation in automotive body assembly. This algorithm make use of measurement points characteristics (assembly process characteristics) and correlation in data to find group of measurement points on automotive bodies. Each group of measurement points is considered as a potential case study, and principal component analysis is applied to each group. The eigenvectors from the principal component analysis is displayed showing patterns of variation on bodies. From these patterns, root causes of variation can be identified.