Quantifying the Environmental Characteristics Influencing the Attractiveness of Commercial Agglomerations with Big Geo-Data
Zhou Huang,Ganmin Yin,Xia Peng,Xiao Zhou,Quanhua Dong
DOI: https://doi.org/10.1177/23998083231158370
2023-01-01
Abstract:Understanding the attractiveness of commercial agglomerations contributes to urban planning. Existing studies focus less on commercial agglomerations, and most directly use environmental supply factors to characterize attractiveness. This study measures attractiveness from the perspective of human demand. Specifically, we build a novel bipartite graph based on big geo-data of human mobility, using node centralities (degree, betweenness, and pagerank) to measure attractiveness. Next, we summarize multisource environmental features such as Point-of-Interests (POIs), land cover, transportation, and population, and use them as inputs to accurately predict attractiveness based on random forest. Finally, the spatial heterogeneity of the effects of these environmental variables on attractiveness is analyzed by multiscale geographically weighted regression. The results of the Beijing case show that: (1) All three centralities show a trend that the urban center is higher than the surrounding area, and betweenness is more reasonable. (2) Random forest can accurately predict attractiveness, with R2 for degree, betweenness, and pagerank at 0.903, 0.846, and 0.760, respectively. (3) The number of shopping POIs, the length of main roads, and the number of bus stops positively affect attractiveness, while the effects of greening ratio and population density are bidirectional. As for the service scope, about 70% of commercial agglomerations have an average service radius of less than 15 km, which is significantly correlated with the Voronoi diagram. Our results can inspire understanding the human–environment relationship and guide urban policymakers in business planning.