On Spatiotemporal Patterns in Urban Violent Crime

"Urban crime, as understood in modern criminology is a sociological, and further a sociogenic phenomenon governed by complex, dynamic processes. Characteristics of the civic, social, and urban environment influence where and when crime events oc- cur; however, past studies often analyse cross-sectional data for one spatial scale and do not account for the processes and place-based policies that influence crime across multiple scales. This thesis builds upon a curated dataset for studying reported crime incidents in context over the last 10 years in the City of Chicago, and (i) applies a Bayesian cross-classified multilevel modelling approach to examine the spatiotemporal patterning of violent crime in the city; and (ii) motivates directions at the intersection of Bayesian computation and data assimilation for supplementing theoretical work in mathematical criminology with observed. We utilise a grid-based and subsequently a stochastic partial differential equations (SPDE) approach to the modelling of violent crime incidents as Gaussian spatial processes and elucidate the computational infras- tructure that enables sparse approximations to solution Gaussian fields of the said SPDE such that Bayesian inference is computationally feasible."