JAILCD2025-037: Analyzing Abnormal Crowd Gathering Behavior During Natural Disasters Using Wi-Fi Handshake Data: A Case Study of Shenzhen's 9.7 Rainstorm

Abstract

Global warming has significantly increased the frequency and intensity of rainstorm and flood disasters, highlighting the critical importance of accurately identifying at-risk populations during such events. Traditionally, this process has heavily relied on expert judgment, which, while valuable, often lacks the precision and scalability required in rapidly evolving disaster scenarios. The emergence of spatiotemporal big data presents a transformative opportunity to enhance the identification of at-risk populations. This study proposes a novel methodology for detecting abnormal crowd gathering hotspots during disasters, leveraging high-resolution Wi-Fi handshake data. An empirical analysis of the September 7, 2023, rainstorm and flood disaster in Shenzhen demonstrates the effectiveness of this approach, revealing the presence of numerous unofficial shelters within affected areas. These findings offer valuable insights for urban emergency management and disaster prevention strategies, providing actionable guidance for the efficient allocation of rescue resources and the scientific planning of dual-use shelters designed to function effectively during both disaster and non-disaster periods.