Tracking World Leaders' Fluctuating Popularity With Big Data
BY Jessica McKenzie | Tuesday, March 25 2014
Every morning, if you so choose, you can wake up to a global world leader's popularity report delivered straight to your inbox, courtesy of the Global Database of Events, Language, and Tone (GDELT). The GDELT World Leaders Index ranks world leaders based on the tone (positive or negative) of global news coverage.
The scale ranges from most positive (+100) to least (-100). This morning, for example, Mamnoon Hussain of Pakistan came out on top with an “average tone” of 3.6749 out of 100. President of Venezuela Nicolas Maduro came in last (out of 56—leaders who did not receive enough mentions in the press within the past two days are excluded) with -8.4533.
Of the 56 country leaders ranked this morning, only four were overall positive. The others, including Barack Obama in 25th place with a score of -2.0393, were all in the red.
After the daily ranking there are timelines for each world leader, showing how their relative popularity—in the media at least—has fluctuated over time.
GDELT explains a bit how “tone” is determined, and the various factors that will influence a leader's score:
Tone synthesizes both emotion expressed about a leader and the emotional context of events the leader is currently involved with or being contextualized within. News coverage of major events tends to mention a head of state by name only when that leader is directly involved in the event, is directly affected by the event, or issues a formal statement or otherwise takes action or is connected to the event. Thus, a head of state which orders the police to violently disperse protesters will be contextualized in the massively negative tone associated with that violence. In a similar vein, if the world condemns a leader who fails to take action to stem violent unrest, that leader would similarly be associated with the resulting strongly negative tone.
The GDELT World Leaders Index is just one facet of a massive data project. The Guardian exaggerates only slightly when they call it, “a big data history of life, the universe and everything.”
Led by Kalev Leetaru, of the University of Illinois, it is a collection of more than 250 million events cataloged and tagged using a 300 word taxonomy. Although it currently only contains events going back to 1979 ("all national and international news coverage from the New York Times, all international and major US national stories from the Associated Press, and all national and international news from Google News with the exception of sports, entertainment, and strictly economic news”), eventually they hope to include events going all the way back to 1800.
This Foreign Policy post describes some of the potential applications of this massive dataset, which include tracking and event predicting political events.
Other cool things by GDELT include the Global Knowledge Graph, which “attempts to connect every person, organization, location, count, theme, news source, and event across the planet into a single massive network that captures what’s happening around the world, what its context is and who’s involved, and how the world is feeling about it, every single day.” It also created a map of the recent violence in Ukraine.
Personal Democracy Media is grateful to the Omidyar Network and the UN Foundation for their generous support of techPresident's WeGov section.