Crime is getting a one-two punch.
Round one was data-driven crime fighting. The NYPD pioneered this approach to combating crime -- and it has spread.
Now, data crunching is headed for round two. Instead of analyzing arrests and crime reports to react to crime, police departments are starting to sift through the data they collect to predict where crimes will happen in the future.
This new approach is called predictive analytics and it's happening now because we're in the age of Big Data.
Smartphones and tablet computers. Cloud computing. Faster, ever more sophisticated analytics. The ability to more cheaply sift through massive amounts of data.
These different kinds of innovation all make it possible for us to collect, combine and analyze in ever nearer to real-time huge amounts of data. Now we can have every scrap of crime-related data imaginable at our fingertips. It's now possible to uncover connections that were buried among disparate bits of information before, and get those insights into the hands of police when they need them.
Using predictive analytics, officers can sift through data, such as the time, dates, and locations, say, of muggings and burglaries, to create models of where similar crimes are likely to happen in the future.
Police departments can also take all types of data, whether it's the locations of ATMs around a city, payday dates for local employers, and even school truancy rates and layer that onto the crime data to unearth unexpected, previously unnoticed trends.
And they can run models using the data being collected to test whether their hunches are correct about what sparks crime.
In Richmond, Va., this kind of modeling was used to prove that the police department was correct in suspecting that violent crime would rise after a gun show was held in that city.
In Memphis, which has a leading predictive policing program called Blue Crush, violent crime is down 26 percent since the new approach was rolled out. Using predictive analytics, for instance, the city analyzed data such as the times and locations of about 5,000 sexual assaults and learned that victims were being attacked when they used outdoor payphones near convenience stores at night. Assault levels dropped once the phones were moved inside the stores.
Predictive analytics can help in other ways. We all know what it's like to feel overwhelmed by information these days, even as it helps us be better informed. Predictive analytics provides a new way for the police to wade through this deluge, to hear the signal out of the noise. It can help them get more done, more effectively.
That's really what's so powerful about this new approach to public safety. Because it doesn't replace traditional police work. It just amps it up by using technology to make it easier for police to do the work they already do. Predictive analytics simply uses the information about the world around us to uncover seemingly disparate connections and hunt out unnoticed trends.
And that's what police do anyway. They study human behavior and use those insights and their experience to catch the bad guys.
To learn more about IBM's public safety, click here
The problems he describes being solved by this technology are also solvable with common sense policing. But the police aren't going to use it for that. This is a tool for targeting entire neighborhoods. They won't be wealthy white neighborhoods.
What we're seeing here is the first efforts in the direction of detecting 'precrime' Minority Report-style. They'll be targeting individuals as soon as they can.