Skip to content
Microsoft Industry Blogs

Police departments are well aware that the mere presence of a uniformed officer or marked patrol car is a deterrent to many kinds of criminal activity, but getting police in the right places before a crime occurs has always been a challenge. Tight budgets often make it difficult to put more officers on the streets, but by applying more powerful analytics to existing data, law enforcement agencies can improve their understanding of crime and make better use of their resources through predictive policing.

Traditionally, policing has been reactive: officers patrol their regular beats or districts and respond to reports of crimes as they occur. But by knowing when and where crime is most likely to occur, police can be positioned where they can do the most good—not just by arriving on the scene after a crime, but by preventing a crime from happening. Of course, this concept is not new; departments have long tried to concentrate their manpower in known high-crime areas. But now, the technology exists to take predictive policing to a new level by tapping into sources of data that are currently going unused.

Today’s law enforcement agencies often have multiple legacy computer systems and databases that are siloed from one another – Computer Aided Dispatch (CAD) systems, records management systems, geospatial information, crime analysis systems, offender management systems, and others contain large amounts of potentially useful statistical data, but they are in different formats and the systems do not communicate. Advances in data analytics and machine learning now allow this disparate data to be analyzed, helping departments pinpoint not only where crime is likely to occur, but when and under what circumstances. Where do people gather after sporting events? What happens when the weather turns warm? How do patterns change when school is out? By leveraging computer models that take into consideration historical crime trends, demographics, climatology, geospatial information, and other data sets, law enforcement agencies can better plan where to deploy their resources.

When predictive policing is mentioned, it inevitably brings comparisons to The Minority Report, the science fiction movie in which people are arrested before they can commit a crime. In reality, however, predictive policing is not about making arrests, but about preventing crime through more effective and efficient resource allocation. The Los Angeles Police Department credits its predictive policing program with contributing to a 23 percent decrease in serious crime over the last five years in the department’s Foothill area – home to more than 250,000 people – including a first-ever “day without crime” on Feb. 13, 2014. Microsoft has partnered with a number of cities, including New York, to develop the Domain Awareness System, a sophisticated solution for crime prevention and counterterrorism that aggregates and analyzes public safety and open source data.

Because much of the technology and data being leveraged for predictive policing already exists, it often requires only an incremental expense to enable predictive policing by the analysis of this data. Technology becomes a force multiplier by allowing departments to better position police officers not only to respond to reported crimes, but where their presence will deter crime. Maintaining a police force is expensive, but the cost of technology continues to diminish. Data analytics is a cost effective way to stretch budgets when money for policing is tight.

Predictive policing will never replace traditional law enforcement, but it will certainly enhance the effectiveness of existing police forces, allowing for an informed and practical approach to crime prevention rather than merely making arrests after the fact.