Several Belgian police districts are collaborating with Ghent University on an innovative big data policing model to predict crime. The BIGDATPOL project uses crime statistics combined with Artificial Intelligence (AI).
"We want to develop evidence-based models for big data policing and we cannot do that without doing extensive research in the field, and therefore working together with investigative agencies and relevant stakeholders," says UGent project leader Wim Hardyns. "There is very little strategic knowledge about crime, a large well-developed database could make a huge difference. Thanks to that data, in the future we can especially estimate where and when crimes will take place."
Police zones participating in the project are divided into street blocks. A relationship is then established with available data. Hardyns: "We also take into account whether there are stations or cafes nearby, for example. In areas with little street lighting, the likelihood of residential burglaries is higher. The presence of escape routes plays a big role. And, also the weather is important, criminals also prefer to stay home when it rains."
