Predictive policing is on the rise in the US, UK and Europe. The technique now faces one of its toughest challenges: the Felony Lane Gang
THEY always choose the line at the bank farthest from CCTV – that's how the Felony Lane Gang got its name. With crimes committed in 34 states, they've withdrawn millions of dollars from banks using cheques and credit cards stolen from cars. A handful of individuals connected to the group have been arrested, but the ringleaders have remained at large for years. Can crime-predicting software finally stop them in their tracks?
That's the hope of police in the US, who have begun using advanced software to analyse crime data in conjunction with emails, text messages, chat files and CCTV recordings acquired by law enforcement. The system, developed by Wynyard, a firm based in Auckland, New Zealand, could even look at social media in real time in an attempt to predict where the gang might strike next.
"We're trying to get to the source of the mastermind behind the criminal activity, that's why we're setting up a database so everybody can provide the necessary information and help us get higher up the chain," says Craig Blanton of the Marion County Sheriff's Office in Indiana. Because Felony Lane Gang members move from state to state to stay one step ahead, the centralised database is primed to aggregate historical information on the group and search for patterns in their movements, Blanton says.
"We know where they've been, where they are currently and where they may go in the future," he says. "I think had we not taken on this challenge, we along with the other 110 impacted agencies would be doing our own thing without better knowledge of how this group operates."
It's not the only system that police forces have at their disposal. PredPol, which was developed by mathematician George Mohler at Santa Clara University in California, has been widely adopted in the US and the UK. The software analyses recorded crimes based on date, place and category of offence. It then generates daily suggestions for locations that should be patrolled by officers, depending on where it calculates criminal activity is most likely to occur.
Kent Police in the UK have been using PredPol for two years. A few months ago, officers were given a patrol location by the system and initially thought it strange – it wasn't in one of the areas most commonly affected by street crime. They visited the location anyway and discovered a distressed woman and a child in public. The woman had been beaten up and the child sexually assaulted.
"The officers managed to take care of them and also managed to apprehend the offender, who was a known and wanted suspect," says Mark Johnson, head of analysis at Kent police.
He says that when the statistical data for the area in question was analysed, officers realised that although it was not as prone to crime as other areas, similar offences had been recorded there in the past. The software did not predict a specific crime, but it predicted that something violent was likely to take place – and it was right.
Targeting which areas to patrol has had a significant effect. Johnson says that PredPol is one of the reasons why the annual number of recorded crimes in Kent has fallen from 140,000 to 100,000 since its implementation.
Part of the enthusiasm for this technology has come from officers burdened by tightening budgets, especially in the US, says David Roberts at the International Association of Chiefs of Police. "There's been real pressure on law enforcement agencies to work smarter, to do more with less and be much more proactive in targeting their scarce resources," he says.
Burglary here
David Wall, professor of criminology at the University of Durham, UK, thinks statistical technology can be highly beneficial but he warns that not all crimes can be predicted – yet.
"Anomalies can happen anywhere and these are not necessarily related to social circumstances, they tend to be related to circumstances that are unveiled at a particular moment in time," he says. "The classic one is a domestic argument that gets out of hand and turns violent. It's very hard to predict that."
Predictive policing software packages are being adopted across mainland Europe, too. In Germany, researchers at the Institute for Pattern-based Prediction Techniques (IfmPt) in Oberhausen have developed a system for tackling burglaries. Precobs works by analysing data on the location, approximate date, modus operandi and stolen items from robberies going back up to 10 years.
Based on this information, Precobs then predicts where burglaries are likely to happen next. This is tightly defined, within a radius of about 250 metres, and a predicted time window for the crime of between 24 hours and 7 days. Officers are then advised to focus their resources in a flagged area.
Precobs has been trialled in the Swiss cantons of Basel-Landschaft, Zurich and Aargau as well as in a number of German cities including Munich and Nuremberg. Michael Schweer, head of analysis at IfmPt, says that the accuracy of predictions so far is about 80 to 85 per cent – meaning that burglaries happened in most of the areas the software predicted.
If you would like to reuse any content from New Scientist, either in print or online, please contact the syndication department first for permission. New Scientist does not own rights to photos, but there are a variety of licensing options available for use of articles and graphics we own the copyright to.
No comments:
Post a Comment