Paper Type

short

Description

Many large cities in the U.S. have a problem with violent crime, some of which is committed by gang affiliates. Those individuals use social media platforms like Twitter to express messages of loss and aggression, which can grow in volume and disseminate quickly, often serving as credible signals to commit an imminent violent crime. These tweets may be useful to law enforcement and community service workers who seek to mitigate violent crime by halting the criminal activity. Thus, this research focuses on exploring the feasibility of automatically ?nding criminal signaling of gang members on Twitter and examining the relationship between this signaling and daily crime increase per city. Content and dissemination features from this analysis and other auxiliary predictors are used to train supervised classi?ers, with the goal of finding criminal signals in the virtual space before actual crimes are committed in the physical space.

Share

COinS
 

Predicting Violent Crime with Gang Social Media Postings

Many large cities in the U.S. have a problem with violent crime, some of which is committed by gang affiliates. Those individuals use social media platforms like Twitter to express messages of loss and aggression, which can grow in volume and disseminate quickly, often serving as credible signals to commit an imminent violent crime. These tweets may be useful to law enforcement and community service workers who seek to mitigate violent crime by halting the criminal activity. Thus, this research focuses on exploring the feasibility of automatically ?nding criminal signaling of gang members on Twitter and examining the relationship between this signaling and daily crime increase per city. Content and dissemination features from this analysis and other auxiliary predictors are used to train supervised classi?ers, with the goal of finding criminal signals in the virtual space before actual crimes are committed in the physical space.