Investigating the spatiotemporally heterogeneous effects of macro and micro built environment on sexual violence against women: A case study of Mumbai

Abstract Sexual violence against women is a major threat to public safety, whereas a well-designed urban environment plays a crucial role in improving public safety and reducing crime. However, the spatiotemporal non-stationarity of the impacts of the macro-level Built Environment (BE) and micro-level Street Environment (SE) on such crimes has been underexplored. Taking Mumbai as a case study, this study employs the crime generator/detractor/facilitator theory to capture the criminogenic roles of land-use functions to describe macro-level BE, while using Street View Images (SVI) to quantify the micro-level SE. Notably, sexual violence against women is classified into four time periods, and Geographically Weighted Regression (GWR) models are developed to capture the spatial and temporal non-stationarity of criminal behavior. The results highlight the varying impacts of BE and SE variables on sexual violence and confirm their non-negligible and complementary roles. Specifically, maternity homes, casinos, cybercafes, and public toilets have been identified as potential hotspots for sexual violence. The complexity of street facades and the presence of retail stores and fire stations (which imply territoriality and surveillance) may contribute to reducing sexual violence. Moreover, the impacts of these variables on crime vary significantly between day and night, from urban centers to suburbs. These findings offer fine-grained insights for urban design and city management, providing decision-makers with evidence-based recommendations to create safer and more women-friendly public spaces.