|Presenters' slides and posters - International Crime and Intelligence Analysis Conference, 25-26 February, Manchester (UK)|
WHAT WORKS CLASSES
27 September 2016
7-10th November 2016
4th October 2016
24th May 2016
6th December 2016
17th May 2016
7th July 2016
12th July 2016
15th November 2016
5th-16th September 2016
ICIAC 2012 Seminar Stream 6A
Abstracts and slides
Predictive crime mapping
Burglars don't choose at random, so why should we? A predictive policing randomised controlled trial
Toby Davies, Shane Johnson, Kate Bowers, and Ken Pease, Department of Security and Crime Science, University College London, and Supt Alex Murray, CI Paul Ditta, Sgt Erica Field, Sgt Clive Baynton and Andy Brumwell, West Midlands Police
Slides: Yet to be supplied by presenter
Research (e.g. Bowers, Johnson and Pease, 2004) suggests that burglaries cluster in space and time with a regularity that makes them (to some extent) predictable. Other work (Johnson & Bowers, 2010) suggests the risk of burglary systematically varies across different types of roads (e.g. main roads and cul-de-sacs). In this talk, we discuss a large-scale operation, currently underway in the West Midlands, for which we have developed and are testing a new predictive mapping tool that draws on the above two observations. The operation, a collaboration between West Midlands police and UCL Jill Dando Institute of Crime Science, is ultimately intended to test the effectiveness of two different crime prevention strategies (target hardening and police patrols) that use the method to inform the deployment of resources. The analysis of such policy interventions is non-trivial and is often undermined by the inability to rigorously evaluate their effectiveness. Such issues include the problem of unequivocally knowing whether an intervention – instead of something else - was responsible for any changes observed. And, for area-based interventions, whether crime was displaced or benefits diffused. We discuss how, by using a randomized controlled trial and a carefully designed set of areas, we address such issues. In this talk we will briefly discuss the predictive method developed, outline the evaluation design, and articulate lessons so far learned regarding the implementation and monitoring of the operation.
Exploring the impact and effectiveness of the ‘Project Optimal’ burglary reduction initiative in Leeds: a spatio-temporal approach
Nicholas Addis, University of Leeds
Keywords: Burglary, Predictive Policing, Crime Prevention, Optimal Forager, Spatio-temporal analysis.
Slides: Yet to be supplied by presenter
The current paper presents the findings from a Masters dissertation project, which explored the implementation of the 'Project Optimal' Burglary Model in Leeds. The model is a form of predictive Policing used to help identify areas at risk of burglary based on previous offending. The Project Optimal Model was implemented in March 2012 as part of a range of initiatives introduced across the city to address the cities highly publicised burglary problem (see Safer Leeds Partnership 2011; BBC 2011, 2012; West Yorkshire Police 2012). Specifically, the model resulted from the city-wide Burglary Reduction Programme introduced in September 2011 by the Safer Leeds Partnership; Leeds' Safer Cities agency (Safer Leeds Partnership 2011; West Yorkshire Police 2012). Based on the Trafford Burglary Model implemented by Greater Manchester Police in 2010, the model is derived from the Optimal Forager theory of behaviour, and previous research that links this theory to offending (Johnson and Bowers 2004; Jones and Fielding 2011). The model suggests that following an initial burglary offence, the risk of subsequent burglaries is increased over a 400m radius across a subsequent 3 week period.
The current paper explores the impact of the Model, through the analysis of spatial and temporal burglary patterns prior to and following the Model's implementation. To achieve this, the project utilised a range of different methodologies, including those used within the Project Optimal model, whereby a series of colour coded buffers are drawn around the preceding 3 weeks' offences to depict dynamic risk over time. Aoristic Analysis was used to derive timebands with which to disaggregate offences and help facilitate temporal analysis of the data. Methods to help establish displacement or diffusion of benefit across both spatial and temporal contexts were also used, in addition to an evaluation of the Model through assessment of its predictive ability.
Three timebands were derived to help facilitate temporal analysis; 'Early', 'Late' and 'Overnight'. The emerging findings were in line with published seasonal effects of burglary; for example where burglary risk appeared greatest during the Late timeband (incorporating afternoon/ evening) prior to the model, where there was decreased daylight hours and extended cover for offenders (see Coupe and Blake 2006). Visual exploration of the data suggested behavioural patterns indicative of an Optimal Forager, and offenders' behaviour appeared to be governed largely by the need to ensure cover and avoid potential detection, in line with Optimal Foraging theory. Diffusion of benefit was indicated, as was both spatial and temporal displacement, albeit to a lesser extent; and the need for further work to explore this in more detail was discussed. The implications of the project for the future of the Project Optimal Model are discussed, in addition to the practical implications of the project for Operational resource allocation within the West Yorkshire Police.
Page last modified on 30 jan 13 17:02