Site Selection Decision Support Tool Using Geographic Information Systems and Multi-Expert Analytic Hierarchy Process

Abstract

Site selection, the process of locating alternatives for new facilities, is a complex and crucial decision faced by growing companies. Organizations often employ time consuming and informal market research techniques, which may fail to capture institutional knowledge or consider all feasible alternatives. Advancements in geographic information systems (GIS) have allowed for analytical methods to be adopted, but current GIS- based methodologies may only be able to study a small area using expensive software, hardware, or data. The goal of this project is to create a decision support tool that can study a large area using open source GIS software and publicly available data, without the use of high-performance computing. The project client is a business that combines an urban winery, a multipurpose venue, and a restaurant into one facility. The company’s site selection problem focuses on finding locations where there is a high demand for their products and services. Requirements elicitation was performed on several experts, and group aggregation techniques were applied to the traditional analytic hierarchy process (AHP) to generate weights for various decision criteria. Data for each criterion was standardized into a consistent scale and then loaded into GIS map layers. A weighted overlay technique was implemented to rank feasible alternatives in map form. Inter- market analysis was conducted using variables that capture an area’s demand for weddings and corporate events, which are the company’s key sources of revenue. Variables that capture demand for the organization’s services include labor availability, existing event infrastructure, and wine consumption in the target region. Intra-market analysis is performed to provide granular recommendations by capturing factors such as crime statistics, accessibility, and proximity to complementary businesses. Recommendations were provided at a “census block group” level of granularity. Sensitivity analysis was performed to test model robustness, and model accuracy was validated through ex post analysis of the firm’s existing locations. Opportunities exist to apply the underlying methodology presented in this project for other companies in various industries to address site selection problems.