Predictive Analytics the New Force Multiplier for Effective Policing
Salt Lake City, UT (PRWEB) April 02, 2014 -- * Field study in five cities demonstrates accuracy of predictive analytics in predicting future crimes
- Tests show CommandCentral Predictive™ effectively identifies almost 30% of next day’s burglaries
- Results are three-times more accurate than traditional hot spotting techniques
- Effects are more efficient use of resources as agencies use prediction areas to determine patrol focus
- A white paper explaining the data is available on PublicEngines’ website
PublicEngines, a provider of cloud-based solutions that facilitate crime analysis, supply actionable intelligence and increase community engagement for law enforcement, schools, and governments today announced the results of a field study looking at the use of predictive analytics to accurately predict future crime.
The results, published in a new paper entitled “Predictive Analytics vs. Hot Spotting: A Study of Crime Prevention Accuracy and Efficiency,” outlines the results from a multi-city test of PublicEngines’ Predictive Analytics product, CommandCentral Predictive. The study set out to measure the accuracy and efficiency of the predictions in forecasting the next day’s crimes versus a commonly utilized hot spotting technique.
Tests were performed by taking a retrospective analysis of crime predictions, and then overlaying the next day’s crimes on the forecasted location for a period of 100 days to test the accuracy of the predictions. The results found that the product’s algorithms accurately predicted the location of the crimes an average of almost 30% of the time. Results highlighted in the paper were for two cities, although the accuracy ratings were similar for all five cities tested.
The data was compared to traditional hot spotting techniques using kernel density estimation methods, which resulted in an accuracy rating of just over 9%, less than a third as accurate as Predictive Analytics.
"We have spoken to law enforcement agencies who have been happy being able to identify 5% of the crimes through traditional hot spotting," said William Kilmer, CEO of PublicEngines. "But when they’ve seen the results of our Predictive Analytics, they’re blown away. They never expected to see such an accurate solution to identifying next-day crime.”
Additionally, the field tests also show the vast improvements in patrol efficiency that can be gained through Predictive Analytics. Using Predictive Analytics, the identified patrol area needed to cover the 30% of the crimes predicted was on third less area than what is needed using hot spots.
“Simply put, agencies that leverage the Predictive Analytics from CommandCentral Predictive, will have a large impact on reducing crimes simply by making sure officers spend more time in the areas that matter most as part of their patrol. Hot spots alone can’t produce those types of results,” said Kilmer.
CommandCentral Predictive is a cloud-based application that analyzes agency crime data on a daily basis and produces targeted, accurate crime predictions that are easy to use by officers as part of their directed patrol plan. The full white paper on the field study and data is available for free download from the PublicEngines Website at http://www..publicengines.com/predictive-vs-hotspotting
About PublicEngines:
PublicEngines develops easy-to-use, cloud-based software that helps law enforcement, government, defense, and other organizations gather, analyze, and share intelligence to improve public safety. Used by more than 2,000 organizations worldwide, our market-leading solutions are accessible wherever you are, and are easy-to-deploy with minimal training and no additional IT burden or expense. For more information visit http://www.publicengines.com.
Public Relations, PublicEngines, http://www.publicengines.com, +1 8018282708, [email protected]
Share this article