Replicability Challenges in Location Analytics

Abstract

The advances in GIS capabilities have been significant over the past few decades, with commercial and open-source software providing relatively easy access to location analytics for even the most novice of users. While many computational methods are indeed accessible and widely applied, this paper focuses on spatial optimization based location analytics available through GIS because of their increased adoption to address a range of social and environmental issues. The significance of this is that insights, management, planning, decision making and policy results are informed by location analytics. Important questions arise, however, about appropriateness, assumptions and validity, especially when user-friendly point-and-click software is involved. Replicability is at the heart of concerns when social and environmental issues are addressed using GIS-based location analytics. Much interest has been devoted to data uncertainty, frame dependency, modifiable areal unit problem and the theoretical assumptions of developed methods, but little attention has been given to definition and implementation details for many advanced location analytics that can be found in GIS. This paper explores these issues as commercial and open-source GIS software incorporate a wide range of optimization based location analytics may face challenges in being replicable, reliable or reproducible.