Characterizing regional importance in cities with human mobility motifs in metro networks

Uncovering higher-order spatiotemporal dependencies within human mobility networks in transportation systems provides crucial insights into urban structure analysis. In most existing studies, human mobility networks are commonly constructed by aggregating all trips without distinguishing who takes which trip. In contrast, we claim that individual mobility motifs – higher-order patterns emerging from individuals' daily trips – serve as fundamental units of human mobility networks. In this study, we propose two network construction frameworks based on mobility motifs to characterize regional importance within cities. First, we consider hidden dependencies within individuals' trips in one day and construct mobility networks based on this mechanism. Second, drawing inspiration from the PageRank algorithm, we hypothesize that individuals assign varying levels of importance to destinations based on their trip intentions. We develop motif-based and motif-wise network construction frameworks. By leveraging large-scale metro data from three cities, Shanghai, Beijing, and Hangzhou, we construct three types of human mobility networks and assess regional importance in urban areas through node importance metrics. Comparative results indicate that the motif-based mobility network outperforms traditional mobility networks, underscoring the efficacy of incorporating human mobility motifs. Finally, we demonstrate that our motif-wise framework significantly improves the accuracy of regional importance characterization.