A spatiotemporal analysis of collective human mobility patterns and crime variations in Baltimore City

Abstract

Researchers have examined the patterns of peoples’ spatiotemporal movements to understand various aspects of our society, including urban crime risks. While prior research has extensively explored individual mobility, collective mobility patterns have received relatively little attention. The growing availability of mobile device location data offers new opportunities to analyze those collective patterns, which used to be challenging to capture. Traditionally, residential mobility and inward flow have been used to study collective mobility, but other measures—such as outward flow—have often been overlooked. Beyond the primary relationship, the spatial distribution of local security measures, such as the intensity of police patrols, may further influence this connection. Temporal variations in the mobility-crime relationship may also arise due to factors like seasonal temperature changes, shifting activity preferences, uneven distribution of holidays, and differences in traveler composition across days of the week. This study explores the relationship between collective human mobility patterns and crime victimization in Baltimore City. It also examines how mobility-crime connections vary across different neighborhoods with intense police activities and during different periods. Using a comprehensive dataset that integrates mobile device location data, official crime reports, and community-level information, the study examines three crime types: homicide, robbery, and burglary. The primary mobility indicators include residential mobility, inward population flow, and outward population flow. Poisson regression with Moran Eigenvector Spatial Filtering (MESF), controlling for ambient population as an exposure variable, is used for the analysis. Key findings include: (1) higher volumes of inward and outward population flows are consistently associated with lower robbery and burglary rates, while residential mobility is positively associated with burglary rates; (2) neighborhoods with more intense police activity exhibit a stronger positive link between mobility and crime; and (3) both seasonal and intra-week (weekday vs. weekend) variations exist in the mobility-crime relationship, though clear seasonal patterns are lacking. Weekday-weekend differences are particularly notable for inward population flow. Those results highlight the importance of collective mobility in shaping local crime risks and call for deeper investigation into the nuanced dynamics between collective mobility and urban safety.