“I’m Prachi Aswani. I’m an undergrad majoring in data science, with a minor in information systems, at the University of Utah.
My research basically focuses on public crisis communication, specifically looking at the COVID-19 crisis. So on Twitter, when politicians and governments in the U.S. were putting out COVID updates, we were looking at how people actually responded to them—not the actual posts, but the comments on those posts—to see how the communication differed and how the public responded to it.
And let me tell you: Looking at social media data is fun and frustrating. You’ll see a lot of censored stuff. You’ll see a lot of sarcasm, but you’ll also see a lot of good points being made in the comments, which adds value to the discussion.
And so I noticed when I first looked at the data, there were two different categories: basically discussion-type responses and reaction-type responses. Discussion-type responses are when people are actually adding value to the post, questioning it or putting out a concern that’s actually valid.
And then emotional-type responses, which were just, ‘Hey, you killed my grandma,’ and stuff like that. So just looking at those two buckets, I think we set out on a mission to find out how politicians or governments can work on their communication strategies, so that it does not create fear or frustration and still gets the point across.
You would be surprised that the proportion of discussion comments is more than the reaction comments, which is not very common on social media, right? But in terms of crisis, we had a proportion where about three-fourths of them were discussion comments, which makes me think that people are actually taking it seriously. Even though there are emotional responses, people still think it’s important to give genuine, valuable concerns or to ask a question that everyone’s thinking about. And because there’s a lot of fear in crisis situations, not a lot of people joke or are just messing around.
We’re aiming to create an API (application programming interface) that you can use in any crisis situation if you have social media data. So if you extract social media data for even a small crisis—like recently, there have been a lot of winter storms happening—and governments or politicians are making crisis updates on social media, we can actually take that data in and you can use our model. It’ll tell you where you’re lacking and give some recommendations for communication strategies.
And so what I hope to take away from Research on Capitol Hill is basically to start at a local level and get those policymakers to see that this is actually beneficial for them. We’re also writing a research paper so that people can take any local crisis that’s happening and analyze all of the comments they’re getting to improve their communication.”
— Prachi Aswani, 2026 Research on Capitol Hill presenter
