GDELT's combination of historical archives and realtime global feeds make it possible for the first time to construct and evaluate intervention strategies in realtime across fields from misinformation to peacebuilding.
Using feeds like the 150-language realtime Web NGrams 3.0 dataset, it is possible to identify emergent narratives in realtime and even connect them to known fact checks and falsehoods using feeds like the Global Similarity Graph (GSG). Yet, once an emergent falsehood is identified, what is the best way to counter it and how can the effectiveness of those mitigation strategies be measured? GDELT's realtime streams, updated every 15 minutes and soon to be updated every 60 seconds with the launch of GDELT 3.0, make it possible to observe the impact of interventions in realtime. Is the narrative decreasing, stabilizing, increasing or evolving to work around those measures? Who are the key influencers who might be best positioned to counter the narrative based on how past falsehoods have been debunked, using historical data?
Similarly, when it comes to implementing peacebuilding and de-escalation strategies, GDELT's historical datasets can help understand which strategies have proven the most effective in a given context and how contextually dependent a given strategy is. In turn, as strategies are implemented, its realtime latent narrative and emotional datasets can help understand whether violence precursors in the form of narratives are evolving positively against violence and whether physical conflict trends are heading in the expected direction.
In other words, GDELT is essentially a live hypothesis and measurement and evaluation (M&E) platform for the planet.
We are excited to be exploring a number of dimensions in this space that will be debuting this fall, so stay tuned!