The team’s NSF Smart and Connected Communities-funded project pairs advanced data modeling with human observation. They developed Community Resilience Information System (CRIS-HAZARD), a mobile app that lets residents snap photos of rising waters and upload them instantly. Those images feed into an AI-based model that predicts flooding hour by hour across Pinellas County. 

“It’s like the Waze of hurricanes,” Guhathakurta says. “People tell us what they see, and that data helps everyone else stay safer.”

The success of Flood Report depends not just on code but on connection. The team’s University of South Florida (USF) partner is involved in managing the community side—training neighborhood “champions” to promote the app, organize town halls, and maintain trust. “Without that local presence, you can’t just drop in an app and expect it to work,” Guhathakurta explains. “You have to earn the community’s participation.” 

The research also revealed social inequities embedded in climate data. “Flooding doesn’t hit everyone equally,” he says. 

“Low-income neighborhoods face the brunt—less elevation, fewer resources, more exposure.” His doctoral student, Mengjue Zhu, is studying how risk perception and flood predictions affect housing prices across those neighborhoods, offering new insight for city planners.

The project’s data includes images from hurricanes Helene and Milton, sharpening predictive accuracy with every storm. The methodology—using sparse but crowd-enriched data—may soon extend to other natural hazards. “The same framework could work for wildfires, tornadoes, or post-storm damage,” Guhathakurta notes. 

“It’s all about integrating community-sourced data with AI tools.” Next steps include expanding the app to more coastal regions and scaling it to support different hazard types. The app “started as a way to understand one place,” he says. “But every coastal community faces this problem now. Our goal is to make this a tool anyone can use.” 

For students in planning and resilience fields, this work is more than data science—it’s human science. It’s about listening to people, coding with empathy, and turning uncertainty into action.*