Criminal social network intelligence analysis with the gang software

Violent street gangs are a major cause of criminal activity in the United States (Bertetto, 2012). In this chapter, we present a new piece of software called GANG ("GANG Analyzes Networks and Geography") that is designed from the ground up to apply new techniques in social network analysis and exploitation to support law enforcement. In particular, we look to enable improved intelligence analysis on criminal street gangs. The software combines techniques from logic programming (Shakarian, Simari, & Schroeder, 2013), viral marketing (Kempe, 2003; Shakarian & Paulo, 2012), community detection (Newman & Girvan, 2004; Blondel et al., 2008), and geospatial analysis (Hannigan et al., 2013) in a usable application custom-tailored for law enforcement intelligence support. This work is inspired by recent work in law enforcement that recognizes similarities between gang members and insurgents and identifies adaptations that can be made from current counterinsurgency (COIN) strategy to counter gang violence (Bertetto, 2012; Goode, 2012; Everton, 2012b). The main contribution of this chapter is the GANG software, which to the best of our knowledge is the first software that combines the aforementioned techniques into a single piece of software designed for law enforcement intelligence analysis. This chapter contains four sections following this introduction. The first describes the system design and implementation of GANG and its various components. The second section provides an evaluation of the software's performance in the assessment of anonymized real-world data on youth gangs. Usage is the topic of the third section, which discusses the Chicago Police Department's experience with GANG. The final section offers some brief conclusions. System Design and Implementation Personnel from the Chicago Police Department described several issues concerning the intelligence analysis of street gangs. We designed the GANG software to meet the following needs. • Ability to ingest police arrest data and visualize network representations of such data - The police data in question primarily consist of arrest reports, which include the individual's personal information as well as claimed gang membership (if disclosed). This data also infers relationships among individuals arrested together. • Ability to determine extent of group membership - While many gang members will disclose their gang affiliation, some will not, often for fear of legal consequences. Hence, to better allocate police efforts and intelligence-gathering resources, it is important to assign these unaffiliated members with some measure of confidence to a gang.