Geodemographic Segmentation: New Methods, New Clusters

Thomas G. Exter, MapInfo Corporation
Ian Mosley, MapInfo Corporation

This paper describes the data, methodological framework, and results of a new geodemographic cluster system for the US. PSYTE U.S. Advantage -- represents a team effort involving extensive processing of Census 2000 and other databases culminating in a typology of U.S. neighborhoods. The paper includes a discussion of geodemographic fundamentals as well as the rational for developing neighborhood clusters in a consumer marketing context. The primary theoretical focus is on the use and interpretation of pattern recognition methodologies to make assignments of small areas (census block groups) to clusters. A variety of databases were used, however, the Census 2000 SF3 data was the primary source. The paper discusses the use of neural network techniques along with hierarchical clustering methodologies to generate the final cluster solution. The paper includes a discussion of results of field tests in which the cluster system was used in a consumer marketing environment.