Tuesday, 29 July 2008

Crime mapping - for those that can read maps

This cow is small. That cow is far away. Small. Far away.....

I always tend to assume that everyone can read maps, plans and charts as well as I can. The old OS inch to a mile with its web of thin brown contours reveals to me a 3D topography as clearly as an aerial photograph; construction plans spring to life in my mind as tangible as a model, and a chart (with a consciousness of current, wind and tide) quickens into a living liquid landscape. An old-school designer I used to work with taught me it was often a mistaken assumption to presume this ability was universal. He used to explain very slowly and carefully to clients what they were looking at. "This is a plan. We're looking down from directly above at the ground floor. Here's the entrance door ...". The key rule of writing a script for a documentary is not to tell the viewer what they're looking at; the key rule for narrating a plan is to tell the viewer exactly what they're looking at.

I've worked with crime maps in London for some time, and they can be as misleading as they are useful. The data on which they are based is taken from the Met's CRIS crime reporting system, often with data from the CAD disturbance call-out recording system. The software then plots shades of crime concentration on a base streetmap - termed hotspotting - in shades that range from deep dangerous reds to benign light primrose. Overall, they show what you'd expect; crime is concentrated in busy town centres, and in other scattered shopping areas or transport nodes. Not greatly telling. Shoplifting, thefts and thefts from vehicles, vehicle offences, drink-associated crime, assaults and disturbances are naturally concentrated in high streets. We know this already.

Now if you can exclude the town-centre crime from the map, and overlay the location of council estates, the thing will provide a revelation. The hotspots are all in shopping areas or tube or train stations within running distance of the council estates.

Burglary is also interesting. A single burglar can be responsible for scores or even hundreds of burglary offences and a typical drug-addicted burglar will generally operate within 400m of his own dwelling; I saw one dramatic hot-spot map for two successive years that showed burglary in one neighbourhood going from puce to primrose, claimed as a major strategic success. The Sergeant explained to me quietly that in reality they'd accidentally caught a prolific burglar, he'd been sent down and the crime rate for this offence plummeted.

And finally, the crime maps will often not gel with the public's experience. This is the same difference as between the British Crime Survey and the recorded crime figures. People may fear crime if there are visible groups of teenagers hanging about, and will expect to see high crime rates. In reality teen nuisance is often actually nothing more than teen presence - they gather for their own safety, not to cause crime.

So crime maps are fine for those that can read them - and have the ability to customise data display and add overlays of other data. Otherwise, they may be little more than a disincentive for the police to record crime.


Blue Eyes said...

Very interesting, but don't you think that publishing the data is a good idea in itself for "freedom of information" purposes?

Newmania said...

My impression is that the Police will use them for thier non stop marketing and in any caser it was only done to neutralise the Conservative Plans and appear active