​West Point's ORCA software knows if you're in a gangS

Researchers at West Point Military Academy developed Organizational, Relationship and Contact Analyzer (ORCA) to help track insurgents in the Middle East. Now several U.S. police departments are using it on criminal gangs.

Software that maps social networks is nothing new – marketers have been identifying influential "taste-makers" for years now to help shape public opinion. The military has found a use for this type of analysis as well, tracking the associates of known terrorists and even pinpointing who leads each cell.

ORCA was initially used to highlight social and family connections between terrorists in Afghanistan. It was known that a tendency toward militant activity was often shared among family members, and that groups of terrorists tend to share certain core beliefs. Software reinforced that kind of analysis, but it could do it in a fraction of the time.

Major Paulo Shakarian turned the software toward domestic gangs, working with the Chicago Police Department to test ORCA's prowess at sussing out gang affiliation. They began by inputting 5,400 arrest records that occurred over three years. Over 1,400 individuals were included. Those who had admitted gang membership were given a 100% probability of belonging to their particular gang. The ORCA analysis then showed who else was likely to part of those gangs, based on who was present at the same arrests and other links.

​West Point's ORCA software knows if you're in a gangS

ORCA's strength comes in part from its ability to make this information clear to the officers using it. Each individual, and each connection, is assigned a weight or a percentage chance of belonging to a given gang. It is also able to identify highly influential leaders within a gang, as well as people who "bridge" two or more gangs. ORCA can even detect substructures within gangs, noting who is on a drug-dealing "corner crew" and which small sets of individuals are more likely to influence the gang's behavior.

The cops in the test program said that they already knew the things ORCA was telling them, which both confirms ORCA's accuracy and brings up the question: what good is it, then? For one thing, it takes less than a minute to run an analysis on an average computer, certainly faster than it would take police to collect and analyze crime and gang data. Understanding gang substructures could be especially useful – a gang without central leadership might be vulnerable to disruption by well-timed street patrols, while a gang with a core leadership group might be taken down by arresting just a few members. And those "bridges" between gangs could make excellent informants.

ORCA can predict patterns of behavior as well, since gang connections can show who is likely to retaliate when a gang member is murdered, and who is likely to be a target. These can be complex network interactions, since gangs are well aware that cops watch more closely when one of their associates is killed. They'll often arrange for an allied gang to make the hit instead.

The ORCA team (and their awesome logo) are working to include geolocation data into their analyses, which could identify gang territory. And while there's a lot to admire about using technology intelligently to fight crime, there's a lot to be wary of, as well. ORCA could easily be aimed at anti-war or anti-tax activists who haven't done anything illegal.