Viridiana Rios and Michele Coscia have created an algorithm they call MOGO (Making Order Using Google As an Oracle) which processes Google data to track cartel activity. "MOGO does the jobs we could never do," Rios told me in a phone interview today. "It reads all of the newspapers that have ever been published in the last 20 years and extracts information about whether and when a particular cartel is mentioned and where that cartel is mentioned. We get the organization, the municipality in which it is supposedly operating, and the year in which the note was published."

That raw data can then be used to show how a particular cartels area of operation has changed over time. For instance, here's how the Juarez cartel grew over the last decade:

And the Zetas:

And here's a picture of how Mexico's various crime organizations have compared in prominence over the last two decades, with the Zetas, Beltran Levya, and Familia groups bursting onto the scene, and Juarez seeming to fade in comparison:

Rios believes these maps give a more nuanced view of cartel activity than existing ones created by private intelligence firms. "Cartels are not operating all around Mexico," she says. On Stratfor's maps it shows every part of Mexico with dominance by one cartel. That is not true. We analyzed about 2400 municipalities and in those 2400 we only found presence of drug cartels in 730."

While the data MOGO gathered may not completely overturn conventional wisdom on cartel activity -- after all, it's drawn from previously published reports -- it seems like a powerful new tool for analysis and it's easy to think of applications well beyond Mexico.