How Mobility and Accessibility Affect Crime Rates: Insights from Mobile Device Location Data

Abstract

This research study investigates the possible correlations between mobility, accessibility, and crime rate. A rich mobile device location dataset including detailed anonymized location traces of the mobile devices observed in the City of Baltimore was combined with the police arrest records to study how mobility and accessibility affect neighborhood safety. The research team first processed and analyzed the mobile device location dataset to obtain measures of mobility and accessibility. These measures differed from the traditional measures in that they were obtained based on the empirically observed location data. The research team then built statistical and machine learning tools to model crime rates at the census-tract levels, using the calculated mobility and accessibility measures, land-use variables, and socioeconomic-related variables as the covariates. Subsequently, the team focused on the correlation of the crime rates with the mobility and accessibility variables. Results indicated that the mobility and accessibility measures can help improve the performance of crime rate prediction. Also, non-motorized travel might be positively related to burglary. The study seeks to inform decision-makers about the transportation-related issues contributing to the lack of safety and offer transportation solutions to crime-related problems, especially in the neighborhoods suffering from high crime rates.