In early 1990, Prince George's Community College (PGCC), in response to declining enrollments, developed an affordable and locally effective geo-demographic cluster system for meeting the college's research and marketing needs. The system, dubbed "PG-TRAK," is based on a model developed 15 years ago as a corporate marketing tool, and involves statistically re-organizing U.S. Census block data according to the natural "clustering" of neighborhoods by stable socioeconomic, cultural, and life-cycle patterns. Life-style clusters, representing demographically inherent segments of the society, could be considered all-purpose sub-markets, each with its own life-style-driven set of attitudes, values and motivations. In developing its customized system, PGCC obtained the complete demographic data set for Prince George's County at the U.S. Census tract level. A cluster analytic procedure was applied to the 90 most relevant demographic variables to reduce them to a marketing-efficient and socially reflective number of clusters (n=24). The system became operational with the cluster-encoding of PGCC's 1985-90 unduplicated credit and non-credit student lists. This three-part manual is intended as an overview of PG-TRAK. Following a brief background and methodological introduction to the system, the manual proceeds with a detailed, cluster-by-cluster data review comprising the cluster "profiles" for each of the 24 geographic areas identified, The manual concludes with a series of supplementary tables comparing all clusters across key variables. (JMC)