Phone: +44 (0) 1273 90 6919 E-mail: contact@openeyecommunications.com Twitter: OpenEyeComms
Predictive Policing: Buzz phrase of the moment or the latest tool for catching criminals and reducing crime in the police toolbox? Either way there’s a lot of noise and focus around predictive policing at the moment.
It’s said that once is accident, twice is coincidence and three times is a pattern (analysts: feel free to disagree here!) and so it is with ‘predictive policing’ at the moment.
The first predictive policing thing to catch my eye recently was the work that Spencer Chainey from UCL’s Jill Dando Institute of Security and Crime Science (twitter: @SpencerChainey) has recently undertaken with the Trafford BCU of GMP exploring how predictive analysis on burglary could be built on and transferred to other crime problems.
Second were comments made by Met Police Commissioner Bernard Hogan-Howe at his recent London School of Economics lecture when he described the need for Met Police technology to ‘move from green screen to iPad’ and talked about having too many systems that produced lists of things that had been done, whilst wanting more agile, predictive, systems that helped him be proactive in reducing crime and catching criminals.
Third was a Policy Exchange (@policy_exchange) event in London yesterday ‘Pre-Crime and Predictive Policing’.
The well attended event, chaired by Professor Roger Graef and with NPIA Chief Exec Nick Gargan on the panel, featured a presentation from Captain Sean Malinowski, Commanding Officer of the LAPD’s Real-time Analysis and Critical Response Division and Professor George Tita, Dept of Criminology, Law and Society at the University of California.
LAPD have been working on predictive policing for seven years and forecasting for four years. They define ‘predictive policing’ as:
‘a place based approach to crime analysis that utilises algorithm-driven crime forecasts to inform decision making to prevent crime’
The essence of the system is a number of geographic 500 foot x 500 foot ‘boxes’ into which crime is mapped and predicted. Capt Malinowski argues that their approach allows deployment and proactivity that is evidence based rather than gut feeling based, stating that ‘It’s forecasting rather than retrospective compstat based, so we’re not just chasing the dots’.
There has been a six months pilot carried out in LAPD’s Foothill Patrol Division and there are plans to roll out the approach to other divisions.
Key question: does it work? Well they are claiming that their results show that using predictive policing is 8% to 16% more effective than existing best practice. However there are a number of issues to be considered before you all rush off to LA to get your predictive policing fix.
The trial is only run Monday to Friday as there are no analysts available on the weekend, and it is run on three crime types only. There are also issues with telling whether officers actually are ‘in the box’ as they are supposed to be and when they are supposed to be, as LAPD cars aren’t GPS enabled. The number of crimes involved in the trial area are also relatively low. However, as both Capt Malinowski and Nick Gargan pointed out, there have been substantial reductions in crime levels over recent years and continuing big reductions year on year are less likely.
Capt Malinowski is of the opinion that cops have too much data thrown at them at pre patrol briefings and that they are overwhelmed by it all, so the LAPD are looking at how ‘just in time’ data push can be used more effectively and they have plans for an iPad app that will be geofenced to particular boxes (i.e. it will trigger and show when the officer – or at least the iPad – is ‘in the box’) and will produce predictive maps of that area for that shift.
The final and most crucial element of the whole trial is the same as with any initiative: officer buy-in. Bottom line: if you can’t convince the cops who are going to work the street and deliver results, all the predictive technology in the world isn’t going to make it happen for you. Culture rules.
Related posts:
Mike, Your comment at the beginning took me back to 1985 and my first term at University. I was delivering some lie or another to explain my second late submission of an essay when Terry Keefe, my tutor said “Once was an aberration, twice is now a coincidence: three times would be a habit.” (Thankfully the habit never developed – but only becasue the clock was reset to zero from time to time).
On the seminar, it was good and interesting, but I feel there are better developed and more long term pilots in the UK that would have been at least as instructive (although we’d have been denied the joy of terms like “Burglarizing,” “Grand Theft Auto” and “Dosage”). The important thing is that we learn from each other and to that end there were some strong connections made with Sean & his team last week.
NJG
Leave your response!