Data Driven Policing

While the data about past crimes works very well in policing current crimes—as Compstat has long proven—using data to predict where it will occur seems more problematic, but this group claims to have done it, as this announcement from Governing notes.

“Once a week on television's "Numb3rs", a mathematical genius helps his FBI brother solve a crime using his prowess in math. In real life, a quartet of mathematicians from the University of California at Los Angeles have created a new mathematical model that helps police understand how crime hotspots form — and what they can do about them. It tells law enforcement whether a police response will eradicate a hotspot or simply displace it by defining a hotspot as either supercritical or subcritical. Supercritical hotspots arise when small spikes in crime pass a certain threshold, and create a local crime wave; subcritical hotspots arise when one large crime spike, such as the presense of a drug den, draws in more criminals. Rigorous policing simply displaces the supercritical variety, but can completely eliminate subcritical hotspots. The UCLA researchers are collaborating with the Los Angeles Police Department in hopes of eventually predicting where subcritical hotspots will form so they can step in before crime erupts.”