Forecasting Civil Strife: An Emerging Methodology

摘要

From the earliest time of recorded scholarship, forecasting civil strife has been the Holy Grail to political theorists. Yet, without actual data and ability to conduct empirical analyses, until the 1960's such analyses were no more than speculation. The advent of high-speed computing along with collection of data on civil unrest allowed political scientists to empirically test their hypotheses. Yet, these analyses did not result in short term prediction due to the lack of real time data. Today the rise of social media has witnessed a radically different methodology in how we can understand, monitor, and forecast incidents of social strife in real time. This emerging methodology, however, requires a multi-disciplinary effort no one even contemplated until recently. This paper presents results of forecasting events of politically motivated violence based on monitoring open source information (Twitter, blogs, newspaper articles) in 10 Latin American countries by a multi-university, multi-disciplinary team of academics, supported by a grant from the Intelligence Advanced Research Projects Activity (IARPA).