Say you’re a Crip gangbanger and want to declare a “Crip Holiday” for you and your homies. You put it out on Twitter, bring a 9mm pistol along with you, and congregate at a local park at 7:30. You get there only to find out that the police are waiting for you. Sounds stupid? Well, it actually happened.
Quoth Mashable’s Radhika Marya:
The New York Daily News reports that police learned the alleged Crips members were using Twitter to call for a gathering at Amersfort Park in East Flatbush, Brooklyn Wednesday. Members of the NYPD’s Gang Unit teamed up with officers from the 63rd Precinct, the borough task force and the Emergency Service Unit, arriving at the park at about 7:30 p.m, police told the Daily News.
While commendable, the bust illustrates a problem with the emerging discipline of law enforcement social media exploitation–we can’t expect everyone to be that stupid. As law enforcement agencies continue to datamine Facebook, Twitter, and other social media to find evidence of crimes, track down fugitives, identify suspects, and gain intelligence on criminal groups, their targets will inevitably learn to cover their tracks. Gang members, violent “black bloc” activists, and other related groups are not supermen and fall prey to the very same privacy mistakes that law-abiding citizens regularly make. But any group hoping to be successful will place a premium on social media operational security (OPSEC). And many already have. For example, many activist groups wary of police surveillance publish extensive guidelines on practicing OPSEC, which can be accessed for free on the Internet.
Police and federal law enforcement and security units seeking to leverage social media have become, in other words, victims of their own success. This isn’t to say that objects of surveillance will go completely analog. Many have grown so accustomed to the use of social media that they are loath to give it up–and simply will better camouflage their usage. But crime and riots originated thousands of years before the first tweet or Flickr photo, and criminals are perfectly capable of organizing themselves without social media. That’s why law enforcement social media exploitation will be one part of a larger data puzzle.
As Alex Olesker argues, the predominant intelligence challenge is micro-targeting with an increasingly overwhelming amount of data:
For example, at the end of 2010, General Cartwright noted that it took 19 analysts to process the information gathered by a Predator drone, and that, with new sensor technology being developed, “dense data sensors” meshing together video feeds to cover a city while simultaneously intercepting cell phone calls and e-mails, it would take 2,000 analysts to manually process the data gathered from a single drone.
On the law enforcement/homeland security level, police will be working with far smaller quantities of data than the military and the strategic intelligence community, but also far less resources. Even when criminals decide to hatch their plots the old-fashioned way, social media can still be used effectively in combination with other sources of information:
Surveillance footage from the 7-Eleven at 13001 Wisteria Drive shows two dozen suspects entering the convenience store en masse, grabbing snacks before leaving. Police posted the footage on their YouTube channel two days later, which has since tallied more than 140,000 views. …So far, police have identified 16 of the 25 suspects and plan to make arrests this week, Starks said. Most are county residents, according to police. The bulk of those positive ID’s came after investigators turned to the in-school police officer for Germantown and pored through Facebook profiles and high school yearbooks.
This sort of Social Network Analysis is a basic part of policing, and social media acts as another element in this time-honored process. However, Alex correctly points out that Hadoop can replicate and automate the more data-intensive aspects of this process (especially when it involves large and complex groups), as well as carry out textual analysis of large data repositories such as chat forums or email inboxes. Even without explicit instructions about capabilities and intentions, Hadoop can identify relationships and point police and federal personnel in the right direction.