Analysis of mobility discrepancies within urban agglomerations using an extended PageRank algorithm in time-varying multimodal networks

Understanding human mobility is essential for transportation planning. However, most existing studies focus on individual mobility prediction or single-mode evaluation, while the time-varying and multimodal features of urban mobility remain largely overlooked. To address these limitations, this study proposed a novel characterization approach that simplifies dynamic mobility networks into static snapshots and incorporates heterogeneous temporal correlations and modal synergies. Based on network characteristics, an extended PageRank algorithm was proposed to evaluate regional importance. Kumamoto (Japan) was used as a case study to validate the effectiveness of the proposed analytical framework. The results suggest that the importance of peripheral areas increases with homebound trips during the evening peak, and static characterization leads to the underestimation of these areas. Furthermore, simplified characterization of interlayer links also results in an inaccurate assessment of peripheral areas. These findings reevaluate the importance of peripheral areas in urban agglomerations, providing novel insights for transportation planning. More importantly, the proposed analytical framework could identify influential nodes in time-varying networks more accurately.