Book Title: Computational Criminology: AI Applications in Forensic Science and Criminal Justice
Editors: Dr. Xavier Louis, Dr. Surbhi Girdhar, Ms. Aswathi Chandran Nair, Mr. Ravi Kumar, and Ms. Nandini Katare
Chapter: 9
DOI: https://doi.org/10.59646/704/9
Author: Anshika Srivastava
Abstract
Serious urban violence is overwhelmingly concentrated in small social networks. Decades of network-analytic research have shown that a substantial majority of gunshot victims and offenders in a typical American city are connected by short paths through co-offending, kinship, and street-group affiliation.[1] Andrew Papachristos’s analyses of Boston, Chicago, and Newark gun-violence networks established the central empirical claim: serious gun violence diffuses through identifiable social networks in patterns analogous to those of infectious disease.[2] In the Boston cohort he studied, eighty-five percent of all gunshot injuries occurred within a single connected component containing only six percent of the city’s population. Similar patterns have since been replicated in Chicago, New Haven, Oakland, and other jurisdictions, with the cumulative finding that proximity in the co-offending network is a stronger predictor of victimisation than residence in any particular neighbourhood.[3] This finding has reshaped the methodological landscape of violence forecasting. The unit of analysis for understanding serious violence is neither the individual nor the neighbourhood but the network position. The relevant networks are not the gangs as named in police gang databases but the empirical co-offending and co-victimisation graphs that emerge from administrative data. And because exposure to violence diffuses along network ties, network position carries predictive information about future victimisation that is not captured by any individual-level covariate[4]. This chapter examines how graph-based machine learning classical social network analysis, embedding methods, and graph neural networks has transformed the empirical study and operational forecasting of urban violence.
[1]“Andrew V Papachristos, “Murder by Structure: Dominance Relations and the Social Structure of Gang Homicide” (2009) 115 American Journal of Sociology 74, 79.”
[2]“Andrew V Papachristos, David M Hureau and Anthony A Braga, “The Corner and the Crew: The Influence of Geography and Social Networks on Gang Violence” (2013) 78 American Sociological Review 417, 423”.
[3]“Andrew V Papachristos, Anthony A Braga, Eric Piza and Leigh S Grossman, “The Company You Keep? The Spillover Effects of Gang Membership on Individual Gunshot Victimization in a Co-Offending Network” (2015) 53 Criminology 624, 631.”
[4]Ben Green, ThibautHorel and Andrew V Papachristos, “Modeling Contagion Through Social Networks to Explain and Predict Gunshot Violence in Chicago, 2006 to 2014” (2017) 177 JAMA Internal Medicine 326, 328.