At the end of August, I travelled to Beijing to present two papers on our recent work in crime informatics at ASONAM 2014, the IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining and a co-located event, FOSINT-SI 2014, the International Symposium on Foundations of Open Source Intelligence and Security Informatics. This builds upon our recent work on social network analysis; in this case using a large dataset of gun and gang crime intelligence from the Greater Manchester area of the UK between 2000-2006.
The abstracts of the two papers are as follows; you can also read final pre-prints via our GitHub repos (here and here):
Measuring UK Crime Gangs
Giles Oatley and Tom Crick
This paper describes the output of a study to tackle the problem of gang-related crime in the UK; we present the intelligence and routinely gathered data available to a UK regional police force, and describe an initial social network analysis of gangs in the Greater Manchester area of the UK between 2000-2006. By applying social network analysis techniques, we attempt to detect the birth of two new gangs based on local features (modularity, cliques) and global features (clustering coefficient). Thus for the future, identifying the changes in these can help us identify the possible birth of new gangs (sub-networks) in the social system. Furthermore, we study the dynamics of these networks globally and locally, and have identified the global characteristics that tell us that they are not random graphs — they are small world graphs — implying that the formation of gangs is not a random event. However, we are not yet able to conclude anything significant about scale-free characteristics due to insufficient sample size.
DOI: 10.1109/ASONAM.2014.6921592
Exploring UK Crime Networks
Giles Oatley and Tom Crick
This paper describes our experiences with three different crime networks in the UK: burglary, ‘gun’ gangs and retail theft. We present an introduction into each of these problems, and highlight some of the issues related to over-simplification of the network analysis.
We also review the term ‘third-generation’ analysis, and provide some insights into achieving this, but also conclude that it can be an extremely computationally expensive undertaking.
DOI: 10.1109/ASONAM.2014.6921658
(also see: Publications)
Digital, Data & Policy 