Dr. Robert Muggah’s typical working day involves drugs, guns, and violence. At the SecDev Foundation in Canada and the Igarapé Institute in Rio de Janiero, his team monitors the social media activity of drug cartels and criminal gangs from San Diego to El Salvador. “The U.S. Southern Command recently announced that gangs and cartels were the biggest threat to the U.S. in the Southern hemisphere, rivaling climate change, jihadism and extremism,“ says Muggah.
Central America has the world’s highest homicide rates and much of the violence is driven by young men in gangs. In fact, of the half a million people who die violently every year, only a small proportion die in war zones. Many more are murdered in cities like Ciudad Juárez or San Salvador which are dominated by gangs and drug cartels. Those same gangs and cartels are also active in the U.S.
“It turns out that the gangs, being young men, are very active on social media,” says Muggah. “They use it to tag (mark territory), they use it to coerce, they use it to recruit, they use it to move product, they use it to communicate directives.” Even when gang members themselves don’t use social media, they are often embedded in communities which share information about gang activity online.
It turns out that social media provides an additional window into gang behavior that traditional research just can’t provide. Muggah hopes that analyzing social media can help policy makers to understand which social policies reduce gang violence in the real world, and which do not. “Our end objective is to find ways to help disrupt networks,” he says.
Muggah’s work uses social media for a variety of purposes from mapping out relationships between cartels and gangs to measuring how their realms of influence change in time and space. It’s a huge undertaking:Thousands of data points are pulled and processed per day—mainly from Facebook and Twitter, but also Instagram, Tumblr, and other platforms. That social media data is integrated with information from local experts and interviews and site visits by researchers on the ground.
Since 2013 Muggah’s foundation SecDev has provided advice to USAID, a U.S. government agency that works to end extreme global poverty, to help understand which of the gang violence reduction programs the U.S. government supports in Honduras and El Salvador are most effective.
Muggah didn’t set out to track gangs and drug cartels; he was initially looking for signs that an Arab Spring-type event could take place in the region. “We set up a series of social media monitoring systems in 2012,” he says, “focused on Columbia, Argentina, El Salvador, Mexico to see what are people getting excited about, what are people getting angry about.” He also documented the government response to popular threads.
The team noticed that there was much less social media activity in Mexico than they expected, given that country’s rather sophisticated Twitter culture. “Part of that was because they were being intimidated by these gangs and by cartels,” Muggah explains. “That’s when we started to encounter the Sinaloa cartel, the Gulf cartel and various gangs who were were essentially involved in coercing citizen bloggers, citizen reporters.“ In parts of Mexico, drug cartels have effectively imposed a media blackout. Only a few brave citizen journalists report on cartel activities. Several of them have paid with their lives.
But if cartels have silenced speech, it’s not because they aren’t actually using the platforms. “Drug cartels implement a wide variety of counterintelligence programs and collect data from social media platforms and blogs as a means of exposing potential threats,” says Marc Goodman, a former LAPD cop and author of FutureCrimes. “They are active monitors of social media. If you say something bad about them they will retaliate. These cartels are equally savvy with their social media campaigns, uploading photos and videos of themselves on Facebook and Twitter in the act of decapitating their victims with chainsaws and machetes. In the same way that terrorist organizations whether it is Al Shabaab or Al Qaeda are clever about creating a narrative online that they want to use to project a particular image, the narcos and the gang members have that same capability.”
So the SecDev Foundation started to follow not only online coercion by Mexican drug cartels but also the social media activities of criminal gangs across Central America. Many of these gangs, like the 18th street gang (M-18) and their main rivals the Mara Salvatrucha (MS-13) are actually transnational.
Both gangs were born in the streets of Los Angeles. M-18 was founded by Mexican immigrants in 1959 and MS-13 by Salvadorans fleeing that country’s 1980s civil war. The gangs’ expansion into Central America was partly fueled by U.S. deportations of tens of thousands of immigrants with criminal records back to their countries of origin. In 2012, the U.N. Office on Drugs and Crime (UNODC) estimated that there were 25,000 MS-13 and more than 27,000 M-18 members in Guatemala, El Salvador and Honduras where they were engaged in extortion, kidnapping, human trafficking, drug and weapons smuggling.
In 2014, USAID had active projects in 215 communities in Central America: community policing programs, crime observatories and 120 outreach centers to provide recreational and educational opportunities for youth in high-crime areas. “One of their challenges is to understand, are the programs they are investing in resulting in positive outcomes?” says Muggah. “Does violence go down? Is there less gang recruitment? Do gang members go back into gangs after being in the program? ” Often, Muggah says, USAID would spend “hundreds of millions of dollars” but didn’t have a good sense of which programs were making the most impact. “We are able to do almost a hotspot analysis of the distribution of violence and gang activity before and after these interventions,” he says.
Muggah calls the type of work the SecDev Foundation is doing with USAID, and others, very large-scale conversation mapping. Social network analysis is used to explore online relationships between known gang members and then enumerate them out to three degrees of separation. Shared interests and content were also used to help establish community membership.
A glossary of terms used by various gangs, including slang for trafficking in specific goods or committing certain crimes, is developed with the help of academics and practitioners working on the ground. “We call this our gang lingo monitor,” says Muggah. “Understanding the language of gangs allows us also better map out their leadership, command and control, their various criminal interests, but also their habits and motivations.”
Earlier research by the SecDev Foundation showed that over 53% of gang members in Honduras and El Salvador did not disable geolocation for Twitter and Facebook postings so online activity can often be linked to a specific place. Prisons are also important geographic markers. “We are tracking gang leaders who call the shots from inside their prison cells (where they should not, in theory, have phones), ” says Muggah, “and ex-gang members outside of them.”
Mobile and internet connectivity was mapped using the location of cell towers, landlines and other structures. OpenSignal’s NetworkStats API, which returns data on download speed, upload speed and other metrics for cellular networks, was also used to gauge their coverage. All this was charted against gang networks to better understand any biases in the data coverage.
Finally, Muggah’s team developed a developed a map of gang-violence reduction programs supported by USAID, the World Bank and host governments, what gangs they are focusing on, how they work and their general catchment area.
The final custom dashboard pulls in social media traffic from social data aggregator Datasift and aligns it according to geographic polygons corresponding to El Salvador and Honduras. “We then have a full-spectrum monitoring system tracking gang movements and activity over time, location, and in relation to specific activities, “ says Muggah. “We could explore specific clusters of keywords to infer activities. We could then correspond these to areas where community security and anti-gang activities were underway before, during, and after the interventions.”
The results of the USAID project are not yet public so Muggah could not tell me which anti-gang interventions appeared to be most successful in Honduras and El Salvador (USAID has performed other studies which showed its projects had a positive impact in general), but he can say what has worked in the past. “From our experience, activities that focus on vocational education, building family cohesion and mentorship seem to generate positive dividends,” he says. “In the long-term, activities promoting early childhood intervention and work with single-headed (female) households are also critical and have a strong empirical track record of success.”
The SecDev Foundation recently started working with law enforcement in California to map out the relationships between gangs and cartels in Central America and their counterparts in downtown San Diego and Los Angeles. “San Diego, as it turns out, is like the gang capital of the United States,” says Muggah. “It’s got the Latino gangs. It’s got African American gangs. It’s got Middle Eastern gangs. It’s got Somali and West African gangs. It’s got everything.”