ABSTRACT
The field of environmental criminology, as concerned with spatial crime patterns and crime theory, has a long history of employing mathematical models and statistical techniques to better identify and understand the underlying factors associated with high or low crime rates in particular areas. The main spatial modelling and analysis methods used in criminology are broadly broken down into: regression, hotspot mapping and simulation. Regression models are a family of mathematical models that estimate the relationships between a dependent variable – usually rates or counts of crime occurrences – and a number of explanatory variables that might include environmental or socio-demographic factors. A common mapping technique that is used to visualize large volumes of crime is kernel density estimation (KDE). KDE maps estimate the overall density of a phenomenon from a set of distinct points and are particularly effective at highlighting crime 'hotspots'.
