Comprehensive archives of regional real-time transportation system data, drawn from public agency fixed and moving detectors and integrated across travel modes, can provide unprecedented opportunities for precise and reliable system performance analysis at relatively low costs. Our access to the state-of-the-art Archived Data Management System (ADMS), a large transportation data archive in Los Angeles, has made possible new research aimed at developing strategies to improve the efficiency and productivity of urban transportation systems. This project, an application of the ADMS, presents a flexible framework to examine the characteristics and explanatory factors associated with intra-metropolitan variation in highway system performance in Los Angeles County. Using one year of highway data and employing three different performance measures that capture network traffic congestion, flow and reliability, we analyze the effects of systematic, random and land use factors on performance variation. We find that performance differs across different types of highway segments, and that population density and accidents are significant factors in explaining these differences. Our study sheds new light on spatiotemporal variations in highway system performance within a large and congested metropolitan area. We underscore the need for investing in regional data archives, and applying them for research and analysis in order to improve planning and system management.