Some of the same social media analyses that have helped Google and the Centers for Disease Control and Prevention spot warning signs of a flu outbreak could be used to detect the rumblings of violent conflict before it begins, scholars said in a paper released this week.
Kenyan officials used essentially this system to track hate speech on Facebook, blogs and Twitter in advance of that nation’s 2013 presidential election, which brought Uhuru Kenyatta to power.
Similar efforts to track Syrian social media have been able to identify ceasefire violations within 15 minutes of when they occur, according to the paper on New Technology and the Prevention of Violence and Conflict prepared by the United States Agency for International Development, the United Nations Development Programme and the International Peace Institute and presented at the United States Institute of Peace Friday.
These predictions can be improved by adding other data from satellites, surveillance cameras and other sensors, the authors of the big data section of the paper said.
The authors were careful to note, however, that data analysis isn’t a one-size-fits-all solution for conflict and researchers should always take the specific nature of a conflict into consideration.
Crowdsourcing initiatives to encourage citizens to report violent behavior by Latin American drug gangs, for instance, may rely heavily on anonymity to keep the reporters safe from retribution, said author Emmanuel Letouze, a Ph.D Candidate at the University of California-Berkeley. A separate crowdsourcing system seeking evidence of electoral violence in Kenya, on the other hand, may be damaged by anonymity, he said, because it would encourage false reports from both candidates’ camps.