• Crime-predicting software helps nab one of the most-wanted alleged car prowlers in Seattle

    By: David Ham


    Seattle - Last May, Seattle police started using software called PredPol, short for predictive policing, that is supposed to predict crimes based on historical crime data. Last month, officers nabbed one of the city's top 10 most-wanted alleged car prowlers because of the software.

    Enous C. Pertee, 42, was arrested inside a garage on the 2900 block of 1st Avenue in Belltown. "The officers got the call and they were able to make the arrest because they were in the right location," said Sgt. Christie Robbin. She said it's too early to say if the software is directly related to a drop in property crimes but says the program has been successful. That's why starting later this month, SPD will start using the software to predict violent crimes throughout the city. "It's the same concept, same strategy with the violent crimes. You're going to have less numbers than property crimes so we'll do that citywide rather than precinct-wide," said Robbin.

     Right now, officers see the crime forecasts at the start of their shifts. The software directs them to 500 by 500-square-foot boxes on a map where a crime is likely to happen. "It gives us an effective tool for deploying officers to the right place at the right time for property crimes," said Robbin.

    Officers patrol the boxes on their downtime. Robbin says the goal is to patrol the boxes around two hours a day. She adds that the more officers patrol the boxes, the more likely crime will likely not happen in the areas. It takes about 1,000 minutes of patrolling to reduce crime in a box.

    Seattle police paid about $70,000 to purchase the software and pays about $40,000 a year for the software subscription. It will not cost any more money to use the software to predict violent crimes.



    Next Up: