Crime prediction pilot to forecast Birmingham's burglary hot spots

West Midlands police is to pilot predictive crime modelling to help direct its officers to areas at high risk of burglaries. West Midlands will begin its trial of the modelling system, which is based on analytical techniques developed by the Jill Dando Institute at University College London, this spring. The pilot, which will see the modelling software create colour-coded maps to indicate likely crime hot spots, is likely to run for six or seven months, according to the force's Superintendent Alex Murray. The modelling software uses algorithms previously developed by the Jill Dando Institute, including one created for Merseyside after the institute analysed some 5,000 burglaries. The analysis revealed "some very specific statistically significant vulnerabilities once a burglary has taken place in vicinity", Murray told Guardian Government Computing: for example, within 24 hours of a burglary the most vulnerable areas are either side of the burgled premises are "up to 400 metres either side and diminishing in vulnerability from the burgled premises". Birmingham is also using two other predictive tools to generate a more accurate prediction of burglaries. One is based on an analysis of roads in the area around a crime and the second, which will be tested in 40 wards, analyses the behaviour of burglars who return to the same location, or are deterred from doing so. "We need to get officers in the right place at the right time, to try and deter burglaries taking place, which means officers patrolling the area, visiting houses and providing security products to particular houses," said Murray. "Clearly we are not going to withdraw services from anybody, but currently we do try and put police officers in places where burglaries are going to take place to try and catch and deter burglars, and this is trying to get smarter in how we do that," he added. This article is published by Guardian Professional. For weekly updates on news, debate and best practice on public sector IT, join the Government Computing Network here .