Remember Tom Cruise’s pre-crime unit in Minority Report? How he used “pre-cogs” to see into the future, predict a crime so he could swoop in and prevent it happening?
We’re not there yet. And “pre-cogs” would likely be an unreliable resource compared to the solidity of Big Data.
One man thinks he’s a small step closer to predicting crime in 2017 using data supplied by the Met Police.
Nicolas Gakrelidz, of New York based data science firm Dataiku set about trying to use crime patterns to predict future trouble spots in London.
Writing in his jargon filled blog , he said: “We might already have an idea of who is likely to commit a crime, but how many crimes would this result in, and what would be the nature of these crimes? This was the kind of information I hoped to predict.”
He compiled all the data he could find – here we’re simplifying massively – cleaned it up, created clusters for geographic purposes and then filled a database, number crunching until he could extract crime predictions for 2017.
Check out his map here .
Each circle represents a community. The larger the circle, the more crime is predicted for the area. The largest circles, some of which can be seen near Southwark, Mayfair, and Soho, represent between 4,000 and 5,500 crimes a year.
Also the more vibrant red the circle becomes the greater predicting increase in crime. The bluer, the lower the predicted crime level against 2016.
“The first pleasant surprise was seeing the number of crimes decreasing,” he said.
However, there are a whole bunch of caveats – it doesn’t reflect real-time events and it’s based on data that quickly gets old so may not be relevant.
He said: "Keep in mind, that these models are built mainly with trends with no control and view on the police strategy or structural change that could significantively change the crime level: in this case these results would be wrong. eg adding new police resources, new mall opening etc."
Keep up to date with all our articles on Facebook