Andrew Schuster

Retail management is constantly looking for new methods to minimise the risk and uncertainty associated with locational decisions. To facilitate these ends, this study proposes a raster based spatial decision support system which models sales forecasting regression equations as continuous surfaces. In an attempt to add structure and uniformity to the modeling process, a methodology was set forth including a taxonomy of Map Algebra procedures which can be used to model a variety of explanatory variables commonly employed in sales forecasting. This methodology was applied to a case study of convenience food stores in the municipalities of Kitchener, Waterloo and Cambridge, where, based upon the proposed methodology, the modeling technique appeared to be both accurate in its representation and offer an increased capacity for decision makers to informed choices. This final statement does however; require further justification through subsequent studies to determine more specific measures of both the accuracy and benefits associated with sales forecast surface modeling.