Constructing and Analyzing Criminal Networks

Abstract

Abstract-Analysis of criminal social graph structures can enable us to gain valuable insights into how these communities are organized. Such as, how large scale and centralized these criminal communities are currently? While these types of analysis have been completed in the past, we wanted to explore how to construct a large scale social graph from a smaller set of leaked data that included only the criminal's email addresses.We begin our analysis by constructing a 43 thousand node social graph from one thousand publicly leaked criminals' email addresses. This is done by locating Facebook profiles that are linked to these same email addresses and scraping the public social graph from these profiles. We then perform a large scale analysis of this social graph to identify profiles of high rank criminals, criminal organizations and large scale communities of criminals. Finally, we perform a manual analysis of these profiles that results in the identification of many criminally focused public groups on Facebook. This analysis demonstrates the amount of information that can be gathered by using limited data leaks.

Framework for Analyzing Twitter to Detect Community Suspicious Crime Activity

“…Also, it has been showing that Twitter can help on predicting crime location using linear regression [12]. Furthermore, crime analysis used node analysis for crime detection and understanding key players of terrorist on the social network [13,14,15]. This section is exploring the related work on crime analysis using social media and particularly in Twitter.…”

“…In [14], the researchers have used similar algorithms used in [13]; however, they have also included Page Rank and eigenvector algorithms in the analysis of social network. The researchers used leaked data from data theft service of Nigerian advanced fee fraud scammer, then they have searched for Facebook accounts related to criminal people, getting their profiles.…”

A Survey of Social Network Forensics

“…In this section, we explain research studies which particularly focus on analyzing the ranking of individuals and criminals (Sarvari et al, 2014;Husslage et al, 2015), importance of links between the nodes (Wiil et al, 2010), and the person successor problem (Spezzano et al, 2013). Sarvari et al (2014) suggest that information regarding the organization of a criminal community can be achieved by analyzing criminal social graph structures. Sarvari et al (2014) aim to construct a large scale social graph from a small set of email addresses of some criminals.…”

“…Sarvari et al (2014) suggest that information regarding the organization of a criminal community can be achieved by analyzing criminal social graph structures. Sarvari et al (2014) aim to construct a large scale social graph from a small set of email addresses of some criminals. By using the Facebook profiles that were associated with these email addresses, they constructed a social graph of 43,000 nodes using 1,000 email addresses.…”

Criminal Network Analysis with Interactive Strategies: A Proof of Concept Study using Mobile Call Logs

“…Emilio et al [14] employed Girvan-Newman algorithm and a variant based on modularity optimization called Newmans algorithm to detect and explore the community structures in the CNs reconstructed from phone call logs. Hamed Sarvari et al [15] performed a large scale analysis with clique based methods to find patterns and substructures of that network based on a publicly leaked set of customer email addresses. However, these studies and early researches somehow neglected the importance of network visualization and the interactions during the analysis process, laying emphasis on aspects related more to statical network characterization.…”