As a form of social movement aiming to effect social change, protests could bring about unintended impacts on all walks of life. In other words, the cost of protests can be incurred by those who might not be protesters. The protests triggered by an extradition bill in Hong Kong since 2019 are no exception. This paper focuses on the impacts on the ridership of the metro system on protest days. It synthesizes and hypothesizes factors influencing the distribution of the ridership changes and conducts an empirical study in the context of Hong Kong to study the possible influences and spatial dependence. It is found that, across metro stations, political orientation (percentage of votes to pro-democracy camp in the 2019 Election of District Councils), law enforcement (permission from the police to protest), land use type (especially for commercial and open space), population age and income, as well as transit/road network characteristics and intermodal connectivity, significantly influence the ridership of metro stations during protest days. In addition, the mixed regressive spatial autoregressive model has higher explanatory power than the ordinary least square model, suggesting the need for a spatial lag and error specification. The results could also have significant implications for policy and planning for operating metro services and managing metro stations before, during, and after social shocks.