West Point team builds tool to analyze gang connections
- By Rutrell Yasin
- Jul 19, 2013
Violent street gangs are behind many crimes in cities and towns across the United States, and law enforcement agencies are constantly honing their intelligence gathering technologies to counter gang activity.
To aid police in dealing with street gangs, a team at the West Point Military Academy has released details of a software program that uses new techniques for crime-based social network analysis.
The program, called ORCA (Organization, Relationship, and Contact Analyzer), uses real-world data acquired from the arrests and questioning of suspected gang members to piece together a gang’s social network structure. The software combines techniques based on logic programming, viral marketing and community detection in an application customized for police intelligence support.
That analysis helps ORCA determine the "degree of membership" for individuals who do not admit to belonging to a street gang. It can also help quickly identify influential actors and help divide criminal ecosystems into key sub-groups, according to a research paper about the software. The authors of the report -- Damon Paulo, Bradley Fischl, Tanya Markow, Michael Martin and Paulo Shakarian -- are members of the West Point’s Network Science Center and Department of Electrical Engineering and Computer Science.
The intelligence tool is based on analysis of real-world police data provided by an undisclosed major metropolitan police department partnering with the West Point team. The police department described several challenges in their analysis of street gangs, and they wanted a software system that could ingest and visualize arrest data.
Though criminal street gangs are decentralized, there are often influential members who may encourage criminal activity that is more violent and risky than the norm, the authors of the paper said. Identifying people in positions of influence could provide law enforcement agents insights into how these organizations escalate their crimes.
The West Point team evaluated ORCA on a police dataset of 5,418 arrests from a single police district over a three-year period. There were 11,421 relationships among the arrests. From this data, ORCA assembled a social network consisting of 1,468 individuals (who were members in one of 18 gangs) and 1,913 relationships. ORCA was able to assemble the network within 34.3 seconds on a commodity laptop.
To prepare the analysis, police data was stripped of personally-identifying information. The team used the “individual record” number of each person arrested and created a social network between two individuals arrested together. The software incorporates network visualization to allow for easier viewing of analysis. The West Point team also designed a user interface and a way to generate a full report as a PDF.
The software draws on research that shows similarities in the behavior of gangs and insurgent groups and that identifies how counter-insurgency strategies can be adopted to counter gang violence.
The team’s next step is to integrate geospatial and temporal elements in the analysis. Throughout the summer of 2013, the team is sending project assistants to work closely with the police in order to identify additional police requirements. Currently the police are using the tools for one district but plan to expand to other districts in late 2013.