Listening to Twitter Conversations: Floods, Extreme Weather, and Climate Change

On April 28th, 2022, we explored the ways that Twitter can be used for data analytics and the important potential of this digital resource in environmental warning systems. We were joined by Jim Moffitt, a Staff Engineer on the Developer Relations team at Twitter, who brought his experience in weather monitoring and flood-warning systems to demonstrate the use of Twitter for real-time communication and data exchange. Our own Dr. Jennifer Marlon, Research Scientist and Lecturer at the Yale School of the Environment, moderated the talk.

Watch the event:

Takeaways include:

  1. More than a decade ago, it was shown that notable weather events caused tweet spikes, generally correlated to the intensity of the event.  These tweet spikes, and the content contained in the spikes, are key data sources for public agencies to listen to. Using Twitter, emergency response agencies like FEMA can crowd-source public information to determine which area is hit hardest and find where survivors are. Twitter’s formula of speed and conversation makes it the perfect platform to inform government agencies of the status of the public during emergencies. 
  2. The public can learn from Twitter’s extreme weather capabilities too. Numerous Twitter bots have been programmed to provide the public with regular updates on local weather, rising water levels, and other climatic variables. For instance, the Schuylkill River FWG Twitter account updates local Philadelphia residents with messages such as “On 02/13/2020 at 11:12 the river level is 2.51 feet” so they know when to expect potential floods.
  3. Twitter’s recently redesigned API now includes a section for the academic study of tweets. Built for graduate students and university research groups, this feature allows researchers to sift through and find signals in huge amounts of data. The full archive of tweets, going all the way back to 2006, is available to those with access for such research. The potentials are endless, and may help to highlight the shared experiences of specific communities.

This event was sponsored by the Environment Data Science at Yale and the Yale Center for Environmental Communication at the Yale School of the Environment.