Social and spatial heterogeneities in COVID-19 impacts on individual's metro use: A big-data driven causality inference

While mobility intervention policies implemented during the early stages of the COVID-19 outbreak had a significant impact on public transit use, few studies have investigated the individual-level responses in metro transit riding behaviors. Using long time-series cellphone big data from frequent metro users in Shenzhen, China, we developed a quasi-experimental interrupted time series (ITS) design to estimate the treatment effects of mobility intervention policies on people's daily shares of metro transit use (SMU). The results indicate that the first-level emergency response (FLR) and the public transit restriction (PTR) policy yielded abrupt drops in SMU of 8.0% and 17.6%, respectively, whereas the return-to-work (RTW) order had an immediate recovery effect of 14.5%. The effect of the FLR is time-decreasing while those effects of the PTR and the RTW are time-increasing. Females and elderly people living in neighborhoods near the city center with low population density and fewer transit stations are more adaptable to policy interventions for reducing SMUs, while the recovery effect of RTW is relatively low for the elderly living in less mixed-use neighborhoods with reduced transit service. These findings can help policymakers design more socially- and spatially-precise and equity mobility intervention policies during a pandemic.