Geographically Weighted Regression Approach to Investigate Spatial Variations in Activity Space

Abstract

This study examined individual activity space focusing on whether and how the effects of the built environment vary spatially across the study region. With the use of the 2012–2013 Northeast Ohio Regional Travel Survey data, a global regression model and a geographically weighted regression (GWR) model were estimated to analyze these effects after controlling for sociodemographics and travel time. The global results indicate that people living in areas characterized by high densities, a variety of jobs and populations, and high access and connectivity generally have smaller activity spaces. The GWR results report better fit than the global model and present spatial variations in the relationships between the six built environment variables and individuals’ activity spaces. Relevant transportation and land use policies, such as transit-oriented development and smart growth, may directly affect the built environment and therefore change individual access and sizes of activity spaces. The spatial variations suggest that the effects of these policies may vary. These findings are of great assistance to policy makers as they point out the target areas for effective policy interventions and infrastructure provisions.