Friday, February 24, 2012

Another Example Of Predictive Policing's Promise

The news outlet had this piece that is worth reading about the Santa Clara County Sheriff's Office use of predictive policing to help reduce crime in their jurisdiction. From the piece:
“The most common time [vehicle and residential] crimes were occurring were Tuesdays and Thursdays between 5pm and 8pm,” says Damon, who works out of a sheriff’s office substation in Cupertino. “We put together hot spots and victim profiles to give officers an idea what to look for. In May of 2010, we started seeing a significant decrease. It was pretty immediate once we got our patrol units in the right place at the right time.” 
As a result, from 2010 to 2011, property crimes in the West Valley patrol area for the sheriff’s office—Cupertino, Saratoga, Los Altos and unincorporated zones that include parts of Los Gatos—dropped 23 percent, according to Damon.
Santa Clara SO's program demonstrates the potential that predictive policing has in combating property crimes.   Of course, I'm really interested in seeing the methodology used by these various programs. Some of them are pretty sophisticated. Others might be using simple statistics to calculate days between hits and the likely times of occurrence along with a hotspot map of recent crimes and a bit of analyst's intuition.

No matter how you get to it, if you can provide your officers with a place to be and some times to be there you can help them be more efficient about driving crime down in the community.

Is your agency interested in predictive policing? What would it take for you to bring this to your agency?


  1. Scott, I have been following your blog and working on a project, as a citizen, in Milwaukee. I come from a product manager background similar to the crime analysis from San Jose. As I began to think more about this, it seems the percentage change of the UCR categories is really not a good way to measure changes in crime. A better method might be using the six sigma approach. This approach recognizes that the event, in this case crime, will always exist. And, it has a 'normal' range of occurring. Consequently, law enforcement can change the range within the 95% confidence level but will never reach a level of zero which appears to be the goal based upon a year to year measurement approach. It is my opinion that when we measure year-to-year percentage of change, it 1) does not recognize the range of crime events (this leads to unfair criticism by residents and the press who may see a small increase which is expected at the 95% confidence level) and 2) it seems to assume that a zero 'defect' can be reached. Do you know any cities that use the six sigma type approach and why OCD would not employ this method vs. year to year change? Thanks

  2. Dave, Thanks for the comments. I have not heard of 6 Sigma being used in law enforcement. Departments usually use more that just UCR to determine their operational effectiveness. Other stats involve Calls For Service, Response Times, Traffic Citations, Arrests, Citizen Contacts, etc.

    We know that we are never going to get crime to zero. It's also common to determine what the "normal level" of crime is so we can recognize anomalous levels of crime.

    Also, most agencies calculate crime rates (per capita) in order to determine how population changes affect reported crimes. UCR crimes are convenient as they are published nationwide so it's pretty easy to see how crime numbers are changing for the region, state or the entire country.


I reserve the right to remove defamatory, libelous, inappropriate or otherwise stupid comments. If you are a spammer or are link baiting in the comments, a pox be upon you. The same goes for people trying to sell stuff. Your comment will be deleted without mercy.