Characterizing the temporally stable structure of community evolution in intra-urban origin-destination networks

TL;DR

This paper introduces methods to identify and analyze the stable community structures within intra-urban origin-destination networks, revealing spatially cohesive groups that persist over time and have practical implications for urban planning and public health.

Contribution

The study proposes the consensus OD network and a stable community decomposition method to quantify and identify temporally stable urban regions, validated with Wuhan taxi data.

Findings

Identified 11 stable groups covering 82.9% of regions

Stable groups are spatially cohesive and often within single districts

Cores have higher POI entropy and more services than peripheries

Abstract

Intra-urban origin-destination (OD) network communities evolve throughout the day, indicating changing groups of closely connected regions. Under this variation, groups of regions with high consistency of community affiliation characterize the temporally stable structure of the evolution process, aiding in comprehending urban dynamics. However, how to quantify this consistency and identify these groups are open questions. In this study, we introduce the consensus OD network to quantify the consistency of community affiliation among regions. Furthermore, the temporally stable community decomposition method is proposed to identify groups of regions with high internal and low external consistency (named "stable groups"), where each group consists of temporally stable cores and attaching peripheries. Wuhan taxi data is used to verify our methods. On the hourly time scale, eleven stable groups containing 82.9% of regions are identified. This high percentage suggests that dynamic communities can be well organized via cores. Moreover, stable groups are spatially closed and more likely to distribute within a single district and separated by water bodies. Cores exhibit higher POI entropy and more healthcare and shopping services than peripheries. Our methods and empirical findings contribute to some practical issues, such as urban area division, polycentric evaluation and construction, and infectious disease control.