Multiple centrality assessment of location preferences of retail and services in Petrópolis, Brazil

Abstract

Abstract This article examines the location choices of retail and service activities in Petrópolis, a medium-sized city in Rio de Janeiro state, Brazil. It uses the Multiple Centrality Assessment (MCA) method to measure centrality indices such as closeness, betweenness, and straightness, which capture the location advantage for each node being how close to all other nodes, being how often traversed by the shortest paths connecting all pairs of nodes, and deviating how much from the straight-line connections in the street network, respectively. The kernel density estimation (KDE) is then used to convert both business locations and MCA values at nodes to raster pixels to facilitate the Pearson correlation analysis. Results indicate that all business types are associated with the three centrality indices with statistical significance, but the association strengths vary across business types. Among the 17 business types, auto mechanical shop, car sales and food retail are most competitive in the locations with the highest centrality measures, and especially in closeness. For other businesses, they tend to value betweenness more than the other two centralities.