Hitachi has unveiled a system that it claims can accurately predict where and when crime will strike. Dubbed Hitachi Visualization Predictive Crime Analytics (PCA), the system takes in an exuberant amount of data from a variety of sources and applies machine learning to help find patterns that humans might not spot.
The technology comes from Darrin Lipscomb and Mark Jules, co-founders of crime-monitoring tech companies Avrio and Pantascene. Hitachi acquired the two in September 2014.
Forecasting crime is far from an original idea although as Jules recently told Fast Company, law enforcement has traditionally built prediction models based on their experience with certain variables and gives them a specific weight. The Hitachi model removes the human bias and decides on its own if there’s a correlation between various data points.
It may seem asinine but a lot of criminals air out their dirty laundry right in for all to see on social media, albeit often using code words and so on. Adding in data points from social media can increase the accuracy of predicting crime by as much as 15 percent.
What kind of data sets is the PCA compatible with? Way more than you’d probably expect. We’re told that it can compensate for variables like proximity to schools, nearby subway stations, 911 calls, gunshot sensors and even the weather.
Hitachi is planning to trial the system in roughly half a dozen cities starting in October.
Images courtesy Fast Company