Retail Gravity Models and GIS University of Utah零售重力模型和地理信息系统犹他大学.pptVIP

Retail Gravity Models and GIS Jonathan Dorwart GIS Analyst / Planner Wikstrom Planning and Economic Consultants April 5, 2007 Retail Gravity Models and GIS Building the Model Implementation Value of Model and Lessons Learned * * Typical market analysis estimates competitive position on the basis of trade areas… 5 min. 10 min. 15 min. Fixed radii and breakpoints… Travel time… Sales data from individuals… Better model the impacts of competition when customer datanot available (basis of trade area has to do with character of development) Additional information for making zoning decisions ordetermining optimum amount of retail in a proposed development Allows a “pro forma” determination of the suitability of retail on a site in a given market Pedagogic in the sense that it lets you visualize competition and opportunities Data availability: Census, local area population projections, sales tax and land use data (including competing retail) readily available Why develop a gravity model in GIS to model retail competition? Identify problem Conceptualize main ingredients/symbolize Utility = Size/Distance Operationalize – Define variables Test (observed v. predicted) Herein lies the problem Evaluation How to use results: Advantages vs. Disadvantages Idealization of consumer behavior In absence of empirical data to articulate consumer decision-making process three na?ve assumptions are made: Psychological trade off between size and distance This tradeoff can be generalized across a number of competing sites (all other things being equal) Little consideration of scale or regional context (Original Reilly model was built as an analogy of regional patterns between cities in Texas) Is the model well-specified??? Conceptual Basis Variations on the Gravity Model Reilly (1931) “Social physics” Huff (1964)“Probability” Estimating the friction of distance using multiple linear regression… Gravity model’s multiplicative nature allows log linear form to be