- Combines sophisticated predictive analytics in an easy-to-deploy cloud-based solution
- Provides detailed predictions directly to officers
- Represents extensive work with agencies to field test accuracy and develop product features
PublicEngines announced CommandCentral Predictive, a highly accurate crime prediction tool designed to deliver daily reports to law enforcement officers to aid in predicting and preventing crime.
CommandCentral Predictive solves a major hurdle in most agencies: pushing actionable information to officers in the street. Recent surveys have shown that one of an analyst’s primary responsibilities is to work directly with patrol staff on their tactical patrol needs. Yet many analysts don’t have the time to support every officer and every beat with detailed analysis and direction on a daily basis. CommandCentral Predictive is designed to close the analyst to officer gap by providing a tool that focuses officers on high-probability crime areas and allows analysts to focus on deeper analytical work.
CommandCentral Predictive is also ideal for agencies without a crime analyst as it automatically delivers accurate, actionable prediction reports on a daily basis in a way that officers can use.
CommandCentral Predictive uses a highly accurate and proven prediction engine to provide agencies a tool to be more efficient and effective in predicting and preventing crime. Officers can use daily prediction reports to help create directed patrol plans to maximize agency resources.
CommandCentral Predictive Key Points
- Highly accurate and proven predictive analytics engine
- Daily targeted area predictions delivered directly to officers
- Deployed in an easy-to-use, easy-to-install cloud-based solution
- Key product features built based on direct agency feedback
- Built with the same focus on simplicity and usability as PublicEngines’ other products
PublicEngines added three new crime categories for CommandCentral Predictive.
The three new crime categories are street crime, theft of vehicle, and theft from vehicle. Each category has been highly requested by law enforcement agencies since the launch of CommandCentral Predictive in late 2013. It brings the total crime categories supported by CommandCentral Predictive to five, including previous categories burglary and property crime.
Each category leverages the ensembled prediction engine developed by PublicEngines for CommandCentral Predictive, which analyses historical data and crime interactions to create an accurate, daily forecast of future crimes by shift.
To validate the accuracy of PublicEngines’ predictive algorithms in predicting the new crime categories, PublicEngines performed extensive testing using live data from actual cities. The results of months of testing showed an accuracy level of 33.5% for street crime, 26.3% for theft of vehicle, and 33.3% for theft from vehicle, with some categories and cities showing higher than 40% accuracy in predicting next day crimes. This is comparable to traditional hot spotting techniques using kernel density estimation methods, which result in accuracy ratings of approximately 9%.