論文誌 (国際) CityFlowFragility: Measuring the Fragility of People Flow in Cities to Disasters using GPS Data Collected from Smartphones
Takahiro Yabe (the University of Tokyo), Kota Tsubouchi, Yoshihide Sekimoto (the University of Tokyo)
The 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing (Ubicomp 2017)
Economic loss caused by natural disasters is increasing in many cities around the world. There is an increasing demand for a method that effectively measures the fragility of people flow to appropriately plan the future investment into infrastructure. Conventional methods measure the fragility of urban systems using infrastructure data such as the road and railway networks. However, these methods are costly to perform, cannot directly measure the disruption on human activities caused by disasters, nor can they be applied for individual disasters. Here, we propose a novel method that quantifies the fragility of cities through detecting the delay in commuting activities using GPS data collected from smartphones. Because commuting activities are daily routines for many people, commuting flow has little day-to-day fluctuation, which makes it an appropriate metric for detecting anomalies and disruption in urban systems. This method can be utilized in any city in the world regardless of differences in network structures or population distribution, as long as people commute on a daily basis. We validate our method in various cities for snowfall and typhoons using real datasets in Japan, and show that intuitive results can be obtained. Our method's reliability is clarified by comparing the results with conventional metrics. We also present useful analyses and applications of CityFlowFragility for urban planning and disaster management.