Exploring regional variations in “socio-spatial” interaction and geographic homophily using location-sharing services data

Abstract

This chapter examines how and to what extent there are variations in sub-regional patterns of socio-spatial interaction using a novel methodology. The methodology uses bipartite network modeling combined with spatial statistical and geographically weighted regression analysis. It provides a statistically robust approach for studying regional variations in the relationship between social and spatial interaction at different distance thresholds. The study applies the methods to the Atlanta metropolitan area using a sample of location-sharing services data. While intended as an exploratory analysis, it does provide some evidence that the association between socialization and location behavior and related distance-decay effects are not uniform in space.