The topological structure of urban roads and its relation with human activities at the street-based community level

Abstract

The topological structure of the underlying streets can help us better understand urban space and human activities therein. As human urban movements are inherently heterogenous in space and statistics, whether or not the network of streets holds a similar degree of heterogeneity worth being investigated. Relying on the graph theory and complex-network thinking, we adopted the street segment analysis-based methods and computed segment-based topological metrics in the downtown of two major cities in China: Beijing and Shanghai. More specifically, we used Flickr photo location data as a proxy of human urban activities and counted the movement flow at levels of both street-based communities and street segments. We measured the heterogeneity of each segment-based metric via the extent of being long-tailed in the rank-size distribution (long-tailedness). We found that segment-based betweenness was most long-tailed and was the best metric for capturing human activities within each community and that neither segment-based degree nor can closeness show a similar extent of long-tailedness and can have a good correlation with the segment-based flow. These findings point to the insight that the positive relationship between street structure and human activities is significantly shaped by their shared heterogeneous nature.