A Police that comes way too early

PUNE, INDIA: Criminals follow patterns. Period! Can that mean anything for IT folks? Ever heard of a technology called 'predictive policing'?

It is something that is based on predictive analytics and functions on the following premise — Criminals follow patterns, and with analytics, law enforcement agencies can help determine where the next crime will occur and sometimes prevent it, as Sameer Murdeshwar, Analyst, ValueNotes Sourcing Practice argues in a blog piece.

In the past, this has been used effectively in other areas such as retail to predict consumer behavior, he adds and elaborates.

Murdeshwar feels that pattern detection works best with large amount of historical data dealing with places and times of previous detected and solved crime. Using this data, a model can be built to factor in all attributes related to specific types of crime such as auto theft, murder and robberies.

Accentuating the gravity of the area, he talks about how across the world, especially in urban regions and cities, crime has been rising for the past few years. A cursory search online reveals terrifying statistics such as a rise in violent crimes in more than half the cities that have the highest crime rates in the US, and a staggering hike of 10 per cent for muggings and robberies in the UK. The increase in crime has been exacerbated by the 2008 global financial crisis, which has led to fewer jobs leading to higher unemployment.

"The solution to tackle this problem is simple — more cops on the beat to patrol the streets. The irony is that the same issue causing the problem is also hurting the solution. Budget cuts have reduced the number of patrol officers. Slashing police funding leading to planned attrition among local law enforcement over the past two years has contributed to a rise in street crime." he stresses.

He cites examples like IBM and HP.

"A forerunner in this space is IBM. As part of its 'Smarter Cities' programme, IBM has been helping agencies such as London’s Metroplitan Police, the Polish National Police and a number of US and Canadian cities detect crime. Their program known as CRUSH — Criminal Reduction Utilizing Statistical History, targets high probability crime areas in cities to allow police to deploy troops more efficiently. IBM acquired i2 , a Cambridge based crime analytics company last year to help build capabilities in this field."

Other companies who have jumped on the predictive crime analytics bandwagon through organic and inorganic means include PredPol and HP. PredPol is a company formed after tests and programs were run successfully on crime analytics by scientists at the University of California — Los Angeles, and Santa Clara University. HP acquired Autonomy for $ 11 billion to build its investigative analytics service to help analyze historical crimes and predict and prevent new crimes.

The results from the use of this service has been very encouraging — In Memphis, serious crime has fallen by 30 per cent and violent crimes declined by 15 per cent in the past five years. In New York City, case closings are 25 per cent higher than the national average and crime has decreased by 20 per cent in spite of a decrease of 3,000 officers. As he cautions, predictive analytics does not promise to replace the effectiveness of police patrolling, but it certainly helps law enforcement agencies with tightened budgets and reduced payrolls.