Daily crime flows within a city

Abstract

Objectives: This paper uses transportation data to estimate how daily spatio-temporal shifts in population influence the distribution of crime over a city's census tracts (CTs). A "funnel hypothesis" states that these daily flows are central for crime concentrations within a city. We present arguments for and against funneling prior to empirical analysis. Methods: A municipal transport agency in a large city in Eastern Canada surveyed 66,100 households about daily trips for work, shopping, recreation, and school. This allowed us to link inflows of visitors to numbers of property and violent crimes for 506 CTs. Results: We find strong support for a funneling effect. Daily visitors have a major impact on distributions over this city for both violent and property crimes. Conclusions: Daily spatio-temporal shifts could be significantly more important than fixed residential factors for distributing crime over urban space.