In a packed Academic Innovation Center classroom, Suhong Li, Ph.D., professor and chair of Bryant’s Information Systems and Analytics department, plays a video of Hurricane Milton’s destruction from early October. Footage of downed traffic signs, uprooted trees, and defunct homes suddenly flood TV screens around the room. Discussing the storm’s devastating impact, the subject matter sets the scene — and mood — for Li’s Day of Understanding presentation: How conversations around extreme weather have evolved from 2010 to 2022, and how the public’s emotions toward the topic have shifted over time.

Based on more than 26 million historical tweets extracted through Twitter API, Li’s research revealed that extreme weather discussions encompassed a range of topics, including the challenges and crises posed by climate change, the escalation of extreme weather events, and their global, social, and emotional impacts.

“Discussions on Twitter predominantly focus on floods and droughts/heatwaves,” said Li, noting that there’s been an increase in these types of conversations over time. “Also, posted images focus on hurricanes, floods, and fires.”

Tapping into individuals’ emotional responses to extreme weather, Li used a language model to classify tweets into four emotions — joy, optimism, anger, and sadness.

“Emotion analysis indicates that sadness, followed by anger, are the most prevalent emotions in discussions about various weather events,” said Li. “Topic modeling shows that conversations about extreme weather events are frequently framed within the broader context of climate change.”