Can big data help us predict where crime will strike?

That might sound like science fiction, but the police wanted to test if by analysing large amounts of crime data, also known as ‘big data’, they could spot patterns in the way criminals behaved. Then they’d deploy their resources in the areas theCOMPUTER predicted crime would strike.

POLICING to prevent crime is nothing new. In Britain, it goes back hundreds of years. But if using big data worked, it would mean humans taking advice on where to police from a machine rather than relying on their own experience.

Trials have taken place across the UK, from Kent to Yorkshire. The resultsSUGGEST that predictive policing models can help cut crimes where perpetrators exhibit predictable patterns of behaviour. In 2011, for example, in Trafford, Manchester,POLICE noted a 26.6% fall in burglaries, compared to a 9.8% fall across Greater Manchester in the same period.

However, Kent Police had a slightly less straightforward experience. It ran a successful four-month trial starting in December 2012, but after rolling out predictive policing across the county in April 2013,RECORDED an increase in crime for the following year. It blamed the rise on a failure to deploy resources effectively and inaccurate crime data.

Rachel Tuffin, director of research at the College of Policing, says there is a strongINTEREST in the potential of predictive policing, but further trials are needed. She explains: “Research shows predictive analysis can identify hotspots more accurately, and separate studies show targeting police patrol and problem-solving in hotspots can reduce crime. Forces in the UK and US are testing the effect of combining prediction with action to remove the causes of crime.”

The Metropolitan Police is currently undertaking the UK’s biggest pilot, assessing three types of predictive software, and is expected to publish findings later in 2015.