Rohil Nataraj - Beyond the Minority Report: The Mathematics of Crime Analysis and Predictive Policing

Predictive policing lies at the crossroads of applied mathematics, criminology, and data science, using mathematical models to forecast crime and guide police deployment. Starting with the basis of criminology, this talk then explores the core quantitative frameworks underlying modern criminal analysis, from regression and spatial clustering to briefly introducing more advanced reaction–diffusion and Epidemic-Type Aftershock Sequence (ETAS) models. These models form the mathematical foundation of systems like PredPol, which seek to identify high-risk “hotspots” through spatiotemporal modelling. The discussion extends to more recent approaches, including network models of co-offending, and risk terrain analysis. While the underlying material is technical, the talk will focus on translating these complex mathematical concepts into accessible, layman terms to ensure clarity for a broad audience. The talk will connect mathematical mechanics like differential equations, stochastic processes, and network theory to real-world decision-making and the critical challenges inherent in predictive modelling. This presentation highlights how applied mathematics contributes to the development of data-driven approaches for understanding and preventing crime.