Finding Street Gang Members on Twitter

Area of Science:

• Computational Social Science
• Criminology
• Digital Forensics

Background:

• Street gang members utilize Twitter for intimidation, broadcasting provocative content, and sharing illicit activities.
• Identifying gang members on Twitter is crucial for law enforcement to gather intelligence on criminal activities.
• The small proportion of gang members among millions of users presents a significant challenge for detection.

Purpose of the Study:

• To develop an automated method for identifying street gang members on Twitter.
• To curate a large dataset of verifiable gang member profiles for analysis.
• To analyze linguistic and visual patterns distinguishing gang members from the general Twitter population.

Main Methods:

• Curated one of the largest sets of verifiable street gang member Twitter profiles.
• Analyzed differences in language, images, YouTube links, and emojis used by gang members.
• Trained supervised classifiers using identified features.

Main Results:

• Established distinct patterns in content shared by gang members compared to other users.
• Developed a classifier that achieves a promising F1 score.
• Achieved a low false positive rate in identifying gang member profiles.

Conclusions:

• The developed method shows promise for automatically identifying street gang members on Twitter.
• The findings can assist law enforcement agencies in crime detection and prevention.
• Further research can refine the classifier for improved accuracy and broader application.