← Latest papers
📄 social science

A Macroscopic Framework to Spatially Model Gang Location

This study proposes a macroscopic framework for spatially modeling gang locations and their interconnectivity using Twitter data, revealing geospatial diversity in cities and small-world properties in state connections to inform policy and future research.

Original author: Ryan J. Roberts

Published 2026-11-19
📖 5 min read🧠 Deep dive

Original author: Ryan J. Roberts

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the author. For technical accuracy, refer to the original paper. Read full disclaimer

For decades, the study of criminal groups has been a matter of looking down at the map, focusing intently on the narrow streets and specific neighborhoods where these organizations operate. Researchers have long understood that gangs claim territory, often marking their boundaries with graffiti or specific neighborhood names, and that violence frequently erupts when these borders are crossed. This traditional view treats gangs as local phenomena, isolated within their own communities. However, a growing body of evidence suggests that these groups are not as confined as they appear. Just as people use digital tools to maintain friendships across great distances, gang members increasingly rely on social media to coordinate activities, issue threats, and build alliances that stretch far beyond their immediate surroundings. The question has shifted from simply mapping a single neighborhood to understanding how these groups connect across an entire country, linking cities and states in a complex web that traditional police reports often fail to capture.

This is the territory Ryan J. Roberts, a researcher at Youngstown State University, set out to explore. Rather than relying on police databases, which are often incomplete or restricted, Roberts turned to the public digital footprint left by gang members on Twitter, now known as X. He treated the platform not as a place for casual conversation, but as a vast, real-time ledger of connections. The process began with a careful search for specific language—slang, acronyms, and phrases unique to groups like the Bloods, Crips, or Gangster Disciples. Once potential profiles were identified, the researcher did not simply accept them at face value. Instead, he manually inspected each account, looking for concrete proof of affiliation, such as tattoos, hand signs, specific colors of clothing, or references to gang media. Only profiles that met at least two of these strict criteria were kept. The next step was to verify where these individuals claimed to be located, discarding any accounts where the location could not be confirmed. Finally, these verified profiles were used to build a digital map, connecting cities and states based on the ties between the members.

The resulting picture revealed a network far more extensive and interconnected than previously understood. The study identified 23 distinct gangs operating across 88 cities in the United States, along with connections to seven international cities. When these cities were linked together, they formed a network of 233 connections. Some cities emerged as major hubs; Los Angeles, Atlanta, Chicago, and Compton stood out as the most connected, serving as central points where members from various locations converged. For instance, a member in Los Angeles might have ties to someone in Compton, but also to individuals in Murfreesboro, Tennessee, or Phoenix, Arizona. These connections were not limited to the West Coast or the South; the data showed that a gang member in Chicago could be linked to peers in Reno, Nevada, or Baltimore, Maryland. While some groups remained tightly clustered in their home regions, others displayed a remarkable ability to spread, with members appearing in small, less densely populated towns like Laurens, South Carolina, and Houma, Louisiana.

When the researcher zoomed out to look at the connections between entire states, the pattern became even more striking. The network included 29 U.S. states and 7 international regions, linked by 272 connections. Unlike the city-level map, which showed some isolated pockets, the state-level map was a single, unified web where every state was connected to at least one other. The structure of this web exhibited what researchers call "small-world properties," meaning that no matter which two states you picked, you could likely find a path connecting them in just a few steps. The longest path between any two states in this network was only four steps away. This suggests that the influence of these groups is not just local but regional and national, with members in places like Georgia and Texas having direct ties to international locations such as Indonesia and South Africa.

The study also highlighted the limitations and nuances of using social media as a source of truth. The researcher acknowledged that an online profile does not always guarantee a physical presence; some users might be "wannabes" pretending to be members, while others might be automated accounts known as bots. To address this, the study applied strict verification rules and discarded profiles that showed signs of being automated, such as an excessive number of hashtags or retweets. Even with these precautions, the data represents a digital reflection of reality rather than a perfect census of every individual. The findings suggest that while online interaction does not automatically mean offline violence, the digital ties often mirror the physical networks that exist on the ground. This is significant because it reveals that gang activity is not confined to the urban centers where it is most studied. The presence of these groups in smaller towns and rural areas, where resources for law enforcement are often thinner, presents a unique challenge that traditional, city-focused models might miss.

Ultimately, this research offers a new way of seeing the landscape of organized crime. By mapping the digital connections between gang members, the study demonstrates that these groups are patterned phenomena that span neighborhoods, cities, and states. The data shows that while some groups remain concentrated in their place of origin, others have successfully diffused across the country, creating a network that is both geographically diverse and tightly linked. This macroscopic view allows for a better understanding of how these organizations operate, how they spread, and how they maintain their influence over vast distances. It suggests that to effectively address gang-related issues, researchers and policymakers must look beyond the immediate neighborhood and consider the broader, interconnected web that ties these groups together across the entire nation.

Drowning in papers in your field?

Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.

Try Digest →