NYPD, Microsoft Push Big Data Policing Into Spotlight

If you ever find yourself plotting a crime in New York City, take heed: thanks to the Domain Awareness System (DAS), a new surveillance tool co-developed by Microsoft and the NYPD, you'll probably be in handcuffs before you've even made it to the getaway car.

And if moving your criminal enterprise to another city seems like a solution, you'd better act fast. The tool will be expanding to other cities soon.

DAS was officially unveiled and demonstrated at a press conference earlier this month. Its capabilities offer real-time aggregation of numerous law enforcement databases and a variety of environmental data sources that include: 3,000 closed-circuit television (CCTV) cameras in lower and midtown Manhattan; 2,600 radiation detectors carried by officers throughout the city; and several hundred license plate readers mounted on police cars and deployed at bridges, tunnels, and streets.

[ To say GPS data tracking is a legal gray area is an understatement. See 7 Facts About Geolocation Privacy. ]

NYPD Commissioner Ray Kelly claimed that the system's power comes from not only the nexus of information but also the system's intuitive visual interface, which orients content in geographic and chronological contexts. Articles have widely compared the user interface (UI) to the one used by officers in the film Minority Report, and though DAS lacks the gesture-based controls seen in Spielberg's film, the system--which Kelly called "one-stop shopping for investigators"--otherwise lives up to the sci-fi allusion.

The network of license-plate scanners, for example, allows the system to compare scanned plates against a variety of watch lists. When the system registers a match, it brings up not only the vehicle's current location but also its past known whereabouts, empowering officers to define travel patterns or deduce the suspect's home base. It also considers all other plates that have ever been scanned in the vicinity of the target vehicle within a 60-second window, allowing officers to determine if a culprit might be part of a larger, theretofore unknown caravan. The result is a more efficient allocation of field resources and responses times so fast that many suspects will be nabbed before they even know they're being pursued.

Entrepreneurial-minded New York mayor Michael Bloomberg praised DAS an example of what the public and private spheres can accomplish together. Such collaborations are becoming a trend in his administration; Goldman Sachs recently invested in the city's jail programs and could potentially profit if recidivism falls.

The partnership with Microsoft includes a novel twist in that New York will receive a 30% commission whenever a DAS package is sold. Bloomberg expressed hope that Microsoft will "make a lot of money" selling the system, as New York's 30% cut could enable the city to not only recoup its investment costs, estimated at $30-40 million, but perhaps even turn a profit. This is the first time the computing titan has engaged in such a profit-sharing model, said Dave Mosher, VP of program management for Microsoft's Vexcel subsidiary, in an email.

Early indications are that Bloomberg might get his wish. Mosher claimed that the "response beyond New York City has been pretty amazing since we went public with this." He wrote that "the market for justice and public safety is real" and claimed that discussions are coming not only from the expected major metropolitan areas but also from smaller cities, as well as private organizations with public safety mandates, such as stadiums and universities. Inquiries, he said, have come from both the U.S. and abroad.

DAS becomes one of the most visible big data solutions employed by increasingly tech-friendly police forces. Recent examples are numerous. The San Francisco Police Department--with the help of the San Francisco Citizens Initiative for Technology and Innovation, whose members include around 90% of the city's tech companies--built a Web-based database infrastructure replete with an app that lets officers access the system and file reports with tablets and smartphones. Bay Area neighbor Santa Cruz effectively reduced burglary rates with a predictive algorithm that Time magazine listed among the Top 50 Inventions of 2012. And London's famously robust jungle of CCTV cameras, which includes facial recognition technology, has become such a ubiquitous talking point that the 2012 Olympics included surveillance-themed mascots. A recent study by the Urban Institute found that camera-based systems effectively reduce crime rates, though it claims that not all geographies or offenses are impacted equally.

Many enterprises outside the law enforcement arena are interested in the tech found in surveillance tools. It's hard to say, for instance, who has more interest in facial recognition and biometrics--the FBI, or Facebook. DAS is likely to join the group of tools that appeal to both public and private concerns. Vexcel VP Mosher said that "DAS was built with enterprises in mind." He noted that multiple Microsoft technologies are part of the system, including SQL Server, FAST Search, SharePoint, Virtual Earth Server, and some .NET framework code--underlying capabilities that provide "the built-in ability to instrument almost all aspects of data and sensor feeds for data integrity or other types of operational events and to automatically send alerts."

Mosher claimed that the NYPD's DAS implementation was customer-specific and that the system can be modified to suit the needs of other jurisdictions or enterprises. Facial recognition, for example, is not supported in the NYPD model, but "this is less of a technology capability issue," wrote Mosher, "and more of a customer requirement issue in accordance with [the buyer's] policies." He also said that scalability is not an issue: "We already have a capability that scales down for bandwidth constrained environments where live video feeds are not feasible--and the intent is to have several 'off the shelf' versions with differing data feeds or sensor inputs that are tailored to smaller justice and public safety customers," he wrote.

Mosher said that DAS costs will depend on the level of customization required, with base products similar to the NYPD system being available at somewhat reduced costs. However, "we are talking about millions [of dollars] for a full-blown system," he wrote.

DAS implementation will likely face some legal challenges. Robert Weissberg, a law professor at Stanford, said in a phone interview that he believes DAS is legal but that alerts from its extremely sensitive radiation sensors might inadvertently reveal otherwise undetectable medical conditions of patients undergoing certain procedures--a potential Fourth Amendment violation.

The issue of tracking citizens, meanwhile, is already moving through the courts. United States v. Jones saw the Supreme Court rule earlier this year that authorities cannot place GPS devices on vehicles without first obtaining a warrant. A decision last week from the U.S. Court of Appeals, however, determined that police do not need a warrant to track suspect locations via pay-as-you-go-cellphones. Tracking individuals of interest is markedly different from tracking all motorists in a geographic area, of course, but these cases demonstrate the ongoing legal seesaw surrounding the use of emerging technologies.

Notwithstanding putative legal hurdles, the future of police work almost certainly involves big data tools. "The bad guys have everything that we do, too," said Bloomberg during the press conference. "And if you really want to worry about security and freedoms, that's the first thing."

InformationWeek is conducting a survey on big data. Take our now. Survey ends Aug. 31.