The Growing Use of Big Data and Analytics in Law Enforcement

Big data has always existed in the federal government, from tax information to population numbers. The term big data essentially means a high volume of data that can be processed for decision making purposes. However, its growing presence coupled with emerging technologies, have turned it into a hot commodity in recent times.

Analytics basically plays a role in evaluating the different variables in big data for that output, solution or prediction. Moreover, a key change in the evolution of analytics is the use of autonomous analytics. Previously, data analytics was used for human decision makers, to evaluate and make a final decision. However, through the use of machine learning and other advances, it is the technologies that can now take the next step and actually make the decision or recommended action on their own.

Given this, it is no surprise that the use of big data and analytics is of special use in the law enforcement community. In fact, increased investment in big data tools and services is seen in contract spending under DOJ from FY 2014 to FY 2016. Growing by nearly a third in that time frame, DOJ dedicated $29M in big data related services and technologies in FY 2014, $36M in FY 2015 and $39M in FY 2016. The largest growth in big data spend under the agency has been in analytics, with 31% growth from FY 2014 to FY 2016. Of the $27.5M spent by DOJ in FY 2016 on analytics, $1.9M was categorized as machine data analytics and $1M in predictive analytics. In FY 2016, FBI spent $15M in big data with an estimated $6M devoted to some form of analytics.

Likewise, DHS spend in big data increased by almost 50% from FY 2014 to FY 2016, with $48M reported in FY 2016. Of that, $16.3M was spent on analytics and $16M spent in machine data analytics. Also in FY 2016, some of the department’s primary law enforcement agencies, CBP and ICE, respectively spent $13.7M and $12.5M, making up 54% of DHS spend in big data that year. Note: all of the above numbers are based on FPDS reported spending from FY 2014 - FY 2016 and Deltek’s filtering through use of specific big data keywords.

The most controversial use of analytics in law enforcement is predictive policing – a term that defines the method in which law enforcement uses data, historical information and analytical models to identify and forecast crime prone areas and entities. For example, Assistant Director Scott Smith at a 2017 RSA conference panel stated that the FBI cybercrime unit is focusing the use of predictive policing in its mission, as reported by FedScoop. In essence, the FBI is attempting to forecast potentially harmful cyber activity through use of predictive analytics to pinpoint a breach or other illegal activity before it actually happens.

Moreover, private entities are boasting of data mining techniques to help law enforcement in targeting criminals such as online sex traffickers. These methods rely on machine learning mechanisms in mining online ads and currency payments in order to discern and identify those participating in the illegal criminal activity.  

The most recent example of big data and analytics use in law enforcement comes with the FBI’s release of the Crime Data Explorer. The portal, built in conjunction with 18F, provides reported crime data in a modernized fashion. Making FBI crime data public is not a new phenomenon, however, the Crime Data Explorer provides the trends and rates of crime across the U.S. by state and type of crime and allows users to upload the raw data for further analytical manipulation. Such information will likely be used by other law enforcement entities and the public alike.

In sum, the use of big data will not disappear from the law enforcement sector, rather, its use will vastly increase, particularly as emerging technologies such as artificial intelligence and machine learning are refined. The big data appeal of saving time and resources will continue as a “doing more with less” mindset resonates across all government faculties, including law enforcement.