This paper presents the research design for an ongoing project focusing on modelling propensity to purchase and market potential estimation on a small geographic area basis through the integration of market and geo-referenced data sources. Specifically, the project seeks to use geographical information systems (GIS) and data mining technologies as key market intelligence software tools for market forecast modelling. Using the automobile industry in Sydney as an example, the research aims to improve the use of spatial information and census data in market research, with the outcomes expected to provide valuable information for decision makers in other industries, as well as for market researchers in general