Chapter Four - Change detection of travel behavior: a case study of COVID-19

Abstract

Travel behavior is important in many fields, such as urban management and disaster management. Since the breakout of COVID-19, many people have changed their preference in travel, which is called travel behavior pattern, to respond to the impact of COVID-19. Understanding when, how, and why people change their travel behavior patterns is significant for antiepidemic and estimating the impact of COVID-19 on human society. However, most current studies ignore that travel behavior is multi-dimensions, and it can be a barrier to understanding travel behavior change. To fill up the vacuum of current research, we used an online Bayesian change detection method to detect individual travel behavior pattern change from big mobile trajectory data. For the low data quality problem caused by various and uneven, we design a novel Monte Carlo data grading framework to assess data quality and filter useable data and thus avoid unreliable results. The analysis result shows Tokyo experienced 6 phases of travel behavior change since 2020, and the change was driven by policies to some extent, especially in the frequency dimension and spatial dimension. Also, the correlation analysis indicates the correlation between four travel behavior dimension dimensions, and the infection number provides us with knowledge about how people will make a change in their travel behavior in the COVID-19 period.