Fellow profile: Andrea Fernández Conde

In Mexico, drug cartels are locked in violent rivalries that profoundly affect the entire country and its population. The history of this ongoing struggle is complex and steeped in political controversy, but some believe that careful analysis of the past may help the Mexican government curb the violence.

This was the guiding maxim for Andrea Fernández, a Data Science for Social Good fellow, during her work with the National Security Council in Mexico. According to Fernández (@feranimal on twitter), the conventional wisdom in Mexico over the past decades has been that drug-related violence has spiked dramatically because of the actions of government.

The story goes like this: the government tolerated the cartels for several decades, but starting with the election of President Calderón in 2006, increased police and military operations against the cartels sparked a new wave of violence against the armed forces, public figures and the general population.

“But that’s a reductionist way of looking at the drug war,” Fernández said. According to her, this perspective is the result of an availability bias born from the migration of violence from rural to urban areas. That is, violent events weren’t really new to the country, but suddenly became much more visible to most of the population.

But how do you investigate a nationwide trend involving criminal organizations which are, to say the least, unlikely to respond to your carefully-designed survey, or to give you their institutional data?

This is where Fernández draws upon her passion for interdisciplinary research and her expertise in mathematics and political science. As part of a team of computer scientists, statisticians and government officials, Fernández helped mine years of publicly-available news articles, blogs, and YouTube videos to track the presence of drug traffickers across Mexico and over time. (This research is available here.)

The team then used this drug trafficker data to analyze the escalating violence in certain regions of the country. They confirmed that drug-related violence is much older than most Mexicans realize, and found that turf wars between drug traffickers are the key driver of recent violence in conflict areas – government crackdowns matter, but are far from the whole story. (That research is available here.)

The similar team diversity among the Data Science for Social Good fellows and mentors is part of what drew Fernández to apply to the program. “I’ve always looked for a combination of what I like to do: programming and understanding topics with data. I’ve always been chasing the interdisciplinary teamwork aspect,” she said.

Fernández is passionate about improving the analytical skills of political scientists, a desire that has remained with her since her college days. While pursuing her Bachelor’s degree in mathematics and political science at the Instituto Tecnológico Autónomo de México, one of Mexico’s top universities, she served as both a teaching and research assistant.

Fernández, who described herself as “skeptical,” has no illusions about having an easy project this summer.

“We have a real challenge here,” she said. “We have a great set of skills, we have these really cool projects, but we need to be careful of how we approach them. It’s important for us to provide feasible solutions that will benefit these groups.”

Time constraints are her greatest fear for the summer, which is duly noted since only 10 weeks remain in the fellowship. “My hope is that we’ll cover a lot and at least make a prototype,” Fernández said. She believes that the greatest success would be in creating novel approaches that capitalize on the latest technology, and which have not been available to her partner government agency in the past.

Fernández’s research on the Mexican drug cartels with left her with a negative impression about the Mexican government’s willingness to support such work. However, she’s not without hope. “Some of my friends made