The realm of justice and crime is a vast arena that contains a great deal of actors and how to approach it is a problem for law enforcement and governments worldwide. In many cases, the better informed law enforcement is the more likely they can act to prevent crime rather than having to hunt down criminals. This was the understanding of Sir Robert Peel in the 1800s when he suggested a preventative police force (patrols) which also gave rise to investigative resources (detectives) because of his work the cosmopolitan concept of the officer on the beat puts eyes and ears on the ground to make crowded places more safe.

The delivery of this criminal prevention strategy was outlined in the brilliant book The Queen’s Peace: the origins and development of the Metropolitan Police, 1829-1979. In many ways, modern society has seen an increasing breach of that peace, and we may be looking at better ways to take it back.

However, those crowded places have become even more crowded and many countries have started to lean towards a panopticon—cameras always watching from poles dissolving privacy—that only tends to move crime from watched places to unprotected places (due to static emplacement) instead of proactive or intelligence-based policing actions.

When I heard about PredPol—a software company that builds on computer science and anthropological research carried out at Santa Clara University and the University of California, Los Angeles—and their Big Data crime-intelligence project that they’re testing out with Los Angeles police it reminded me of the step from chasing bad guys after they committed the crime to preemptively patrolling to catch criminals before the act.

According to an article in the MIT Technology Review, PredPol was used by an L.A. precinct called Foothill division to predict hotspots of particular types of criminal activity and used that to inform their patrols on where to concentrate and the beat officers as to what to look for. As a result, they discovered that the software’s predictions seemed twice as likely to be effective than standard analysis.

“We are seeing a tipping point—they are out there preventing the crime. The suspect is showing up in the area where he likes to go. They see black-and-white [police cruisers] talking to citizens—and that’s enough to disrupt the activity,” the article quotes Sean Malinowski, a police captain in the Foothill division, as saying.

The inputs are straightforward: previous crime reports, which include the time and location of a crime. The software is informed by sociological studies of criminal behavior, which include the insight that burglars often ply the same area.

The system produces, for each patrol shift, printed maps speckled with red boxes, 500 feet on each side, suggesting where property crimes—specifically, burglaries and car break-ins and thefts—are statistically more likely to happen. Patterns detected over a period of several years—as well as recent clusters—figure in the algorithm, and the boxes are recalibrated for each patrol shift based on the timeliest data.

Jeff Brantingham, a company cofounder and UCLA anthropologist, said: “The challenge, and what is really hard from the point of view of the crime analyst, is how do you balance crime patterns on different time scales. That’s where the algorithm has the edge, sifting through years of data.”

In the two tests, it appears that using the algorithm has helped reduce property crime by 25% in the targeted areas.

So far, the technology looks extremely promising but with only two tests, it’s hard to say that we’re seeing the totality of the potential outcome. One thing that we need to look out for is that while this is predictive moreso than reactive, it still will tend to cause crimes to move from place to place as criminals discover police presence increasing in any particular region.

Big Data enables a birds-eye-view to inform the block-by-block strategy

Big Data contexts could be used for such an analysis by looking at the statistical likelihood of connections between different type of crimes, the geography of an area, the time of year, or even the weather. Hot spots might crop up for immediate attention; but knowing how a particular hot spot reacted the last time a police presence increase happened there could give officers a chance to catch the consequences of their actions as well.

For example, with enough data delivered into a Big Data algorithm that does crime analysis could also predict not just where more crimes of what type might occur in an area—but also could give better intelligence on where the police could add more beat officers in order to catch outflow or crime fleeing from their activity.

This way, instead of simply moving the geographic distribution of the crimes, the police can proactively make it more difficult for crimes to be committed overall using Big Data and expert agent intelligence. It’s the same way that we can use Big Data to enable better cybersecurity.