Predictive Crime Analytics
Crime Mapping is not a new concept for practitioners of law enforcement. The production of crime maps to identify hot spots or high concentration of crimes within an area has been standard practice for investigators since the mid 19th century, where police departments could better visualise the categories of offences, the frequency and density of crimes by placing coloured pins on maps.
However, with the advent of Geographical Information Systems (GIS), crime mapping has experienced a surge in interest. Over the past ten years police departments and university analysts have explored the potential applications of using GIS to quickly map and visualise large data sets which would have been difficult to analyse without automation. Police and crime reduction agencies are using these systems to aid intelligence development, criminal investigations, crime prevention, performance improvement, information sharing, and crime reduction.1
GIS forms the basis for the recent developments in Predictive Crime Analytics (PCA). PCA is the use of algorithms and analytic techniques to analyse crime data in the context of a range of data from the physical environment, to attempt to predict and anticipate criminal activity in order to prevent it.2 This often includes geographic data and development of PCA systems is closely allied to the use of GIS data.
The majority of PCA is based on the communicability of certain crimes, where one event may be the precursor of another, as well as the idea that certain locations are vulnerable to repeat victimisation. By assessing known causal relationships between past crime events, analytics can inform the police with a prediction of the next likely location.
PCA has been taken up with enthusiasm in the US, where IBM’s SPSS Modeller and Predpol have claimed successes with crime reductions in Memphis (30% reduction in serious crime overall) and Santa Cruz (number of burglaries declined 19%).
In terms of return on investment (ROI), IBM has also claimed that the average annual benefit for the Memphis PD’s adoption of SPSS equates to $7,205,501.3With budget cuts and savings targets effecting Police expenditure and resources, the adoption of PCA systems or tools could prove to be a timely and shrewd investment.
With this in mind, AP Benson was commissioned by the Welsh Government ICT Exploitation Unit (within DETS) to undertake research into the feasibility of developing Predictive Crime Analytics (PCA) activities in Wales. The ultimate aim of this activity is to exploit existing knowledge, capability and expertise in PCA in Wales or to develop these through the development of PCA initiatives.
AP Benson's Managing Director, Tom Girn, said "We're delighted to have been selected by the Welsh Government to research this fascinating field of study. The implications for crime prevention in the UK are immense, as are the potential economic benefits for police constabularies using these technologies and techniques.”
AP Benson is a management and economic development consultancy with offices in Cardiff, London, Birmingham and Reading. The firm assists clients to solve business and organisational problems using innovative thinking and leading edge tools and techniques. AP Benson consultants work in a range of fields including dynamical and financial modelling, economic feasibility, market and business research, project evaluation, business development and IT Transformation.
Gerard Fannon: Junior Researcher
2 Susan C. Smith (2011), Predictive Policing, Shawnee, Kansas Police Department.
3 Nucleus Research, ROI Case Study: IBM SPSS, 2010