Skip to main content Skip to secondary navigation
Main content start

Hackathon focused on equity, prediction, mitigation of wildland fires

Students harnessed the power of big data with the guidance of faculty and industry mentors to propose groundbreaking solutions to wildland fire issues in the Stanford Doerr School of Sustainability’s Big Earth Hackathon.

Collage of people playing role-play game about wildfire preparedness
Snapshots of the The Resilient Futures Playset,” a role-playing board game for preparing for wildfires and planning recovery efforts. The project was created as part of the 2023 Wildland Fire Challenge. (Image courtesy of Derek Fong)

For the 2023 Big Earth Hackathon, a cross-campus competition where students were challenged to innovate solutions to planetary problems, teams confronted the pressing issue of wildfires. During spring quarter, over 60 students participated in the hackathon theme, “Wildland Fire Challenge,” to confront problems related to wildfire prediction and mitigation, and issues related to equity and access to information. 

The Big Earth Hackathon at the Stanford Doerr School of Sustainability is a program adopted from the School of Earth, Energy & Environmental Sciences (Stanford Earth). This year marked the fifth iteration of the initiative, which is led by instructor Derek Fong, a senior research engineer in Civil and Environmental Engineering. 

The first-place winning team of Jierui Fang and Ayisha Jackson, master’s students in Design Impact, particularly impressed the audience and judges with their innovative project, “The Resilient Futures Playset,” a role-playing board game meant to help people collaborate and appreciate different perspectives when preparing for wildfires and planning recovery efforts, and ultimately enable collective decision-making. 

To facilitate consideration of issues surrounding wildland fires, which are exacerbated by climate change, students could participate in the hackathon’s companion course, which featured lectures from faculty and industry experts. They were given eight weeks to think carefully about serious wildland fire problems and propose full-fledged solutions and prototypes. Teams collaborated with industry partners, faculty, and stakeholders to generate relevant solutions to urgent problems. 

“The students really took to heart the different challenges presented by the guest lecturers,” Fong said. “I was especially impressed with how teams used state-of-the-art technology and data science to inform forest management, consider the needs of underrepresented communities, and empower people to make wise decisions.”

Students presented their proposed solutions at a Project Expo on June 2. Seth Schalet, CEO of the Santa County FireSafe Council, attended the expo and commented on the diversity in the types of projects displayed, adding he was “so impressed with the passion and depth of knowledge the students displayed.” 

For the second-place prize, judges selected a project that uses GPT, the artificial intelligence powered chatbot, to provide advice and information in any language based on individual concerns. The project’s lead developer, Sustainability Science and Practice graduate student Thunder Keck, said the tool was intended to help improve access to information among marginalized community members in areas threatened by wildfires. 

The team behind the third-place project, Jesse Moonier (CEE), William Ogden (CEE), and Hannah Young (Earth Systems), built a tool in collaboration with the U.S. Forest Service to evaluate tree mortality risk using low- and high-resolution satellite imagery. Their project demonstrated that more granular imagery provided detailed information about tree mortality but did not provide as timely a warning of impending mortality in comparison to the coarser resolution data.  

Other noteworthy projects included a crowdsourced early wildfire identification tool that can be adapted to different localities, a laboratory experiment showing how firebreak spacing influences firebrand propagation (which won the Best Use of MATLAB Prize), a machine learning model to predict landslides based on wildfire and precipitation history, a wildfire preparedness tool based on real-time information, and a demonstration of a building material similar to particle board that can be made from the brush and small trees targeted in fuel thinning operations. 

“I was impressed with the ability of the students to identify key pain points in the wildfire space,” commented one of the judges, a data scientist from PG&E, the Pacific Gas and Electric Company. “The creative solutions presented highlighted the advantage of having interdisciplinary project teams. The diversity of solutions, from GPT implementation to fluid dynamics to game design, was exciting to chat with students about.”

The Big Earth Hackathon was supported by the Stanford Bits & Watts Initiative, the Bill Lane Center for the American West, the Stanford Geospatial Center, the Sean N. Parker Center for Allergy & Asthma Research, the TomKat Center for Sustainable Energy, Convective Capital, Mathworks, and Microsoft. The judging team was composed of Stanford faculty and staff and an industry expert.

Explore More