Growing pains for global monitoring of societal events.

T here have been serious efforts over the past 40 years to use newspaper articles to create global-scale databases of events occurring in every corner of the world, to help understand and shape responses to global problems. Although most have been limited by the technology of the time (1) [see supplementary materials (SM)], two recent groundbreaking projects to provide global, real-time “event data” that take advantage of automated coding from news media have gained widespread recognition: International Crisis Early Warning System (ICEWS), maintained by Lockheed Martin, and Global Data on Events Language and Tone (GDELT), developed and maintained by Kalev Leetaru at Georgetown University (2, 3). The scale of these programs is unprecedented, and their promise has been reflected in the attention they have received from scholars, media, and governments. However, they suffer from major issues with respect to reliability and validity. Opportunities exist to use new methods and to develop an infrastructure that will yield robust and reliable “big data” to study global events— from conflict to ecological change (3). Automated event coding parses individual sentences into SUBJECT VERB OBJECT format and categorizes the action using a framework like CAMEO (Conflict and Mediation Event Observations). So a statement like “Secretary of State John Kerry complained about Russia’s support of Syria’s Bashar al-Assad” would be coded as US GOVERNMENT/DISAPPROVE/RUSSIAN GOVERNMENT. This can be refined into a numeric level of hostility or cooperation by using scales like the Goldstein Score. Whereas CAMEO focuses on categories for international and domestic conflict, similar frameworks could be developed for almost any kind of interaction in news media (e.g., transactions between businesses or debates over scientific findings). Uses for the resulting data have been manifold. Hand-coded and automated event data have been used to anticipate conflict escalation (2). When combined with statistical and agent-based models, ICEWS claims a forecasting accuracy of 80%. GDELT has been used to track, e.g., wildlife crime and the rise of hate speech following the U.K. Brexit vote. There are several challenges in the current approach. First, the focus on sentences removes a great deal of context. Event occurrences do not neatly partition into sentences. This lack of context, for example, often fails