Duke Students Develop Data Savvy While Tackling Climate Challenges

Undergraduate student Rema Yasir Albarakati discussed her team poster, “Energy Transition During Energy Crisis,” at the Poster Session.

More than 30 students participated in eight project teams in summer 2025.

Over the course of ten weeks this summer, students in Duke University’s Climate+ program used data science applications to research climate challenges and potential solutions. They studied topics like extreme heat stress, flood hazards, electricity access and even Duke’s own greenhouse gas data system.

Climate+ offers students opportunities to take part in small research teams as part of Duke’s Data+ summer program. Undergraduate students work in teams of two to five with a graduate student project manager and faculty leads to collect, analyze and/or visualize data to contribute to interdisciplinary climate research. The program allows students from a range of disciplines to become familiar with data analysis in various forms, such as machine learning techniques, spatial analysis, backend and front-end development and remote-sensing technologies.

For summer 2025, Climate+ teams included:

• Tracking Aquatic Insect Emergence Using Machine Learning
• Energy Transition During Energy Crisis
• Duke University Greenhouse Gas (GHG) Data System Update
• Mapping High-Stakes Coastal Zones
• Adapting to Warmer Temperatures at the City-Scale
• AI-Driven Mapping of Forest Biodiversity Using Remote Sensing
• Forecasting Heavy Rainfall in the Southeastern U.S. Using Deep Learning
• Automated Flood Modeling and Risk Assessment for Milwaukee River Communities

Research from this summer is well on its way to informing climate solutions. One project group worked to develop an interactive dashboard that identifies vulnerable coastal regions, determined by an array of climate and socioeconomic indicators. Effectively synthesizing vast amounts of data, the final dashboard will be the first global coastal social-environmental atlas, offering policymakers, decision-makers and researchers a comprehensive view of coastal conditions.

Undergraduate student Ethan Cho (left) and Data Science Fellow Leonard Eshun (right) presented their research on “Mapping High-Stakes Coastal Zones” at the Plus Programs Poster Session in Gross Hall.

“Through Climate+ I was able to examine the intersection of climate science and technology and came to understand how data science can effectively be used to improve environmental decision-making,” said Duke undergraduate student Ethan Cho.

Another Climate+ team created a modeling approach to estimate heat stress in urban areas. By exploring land cover and climactic datasets from over 60 cities, the students set out to build a generalizable model that could predict urban heat even in regions with limited or no data. Their findings show potential to bring a welcome change from many current urban heat island models that are developed on a city-by-city basis.

Beyond the development of future data tools, Climate+ is opening doors to expand the data science workforce. Duke undergraduate Ella Tallet said she was curious about climate analytics at the summer’s outset, and her time on the urban heat project brought clarity for her path forward. 

“The Climate+ program made me confident in my desire to pursue a long-term career in this realm, and I’m looking forward to applying to jobs related to data and research this fall,” said Tallet.

Since the summer 2022 launch of Climate+, more than 120 students have contributed to 29 interdisciplinary project teams spanning ecology, biology, engineering, environmental science and more.

Climate+ students visited Duke Forest to examine tree canopies for comparison with remote sensing technologies data used by their research team to enhance forest biodiversity monitoring.

Like all students participating in the broader Data+ program, Climate+ students have opportunities to learn from researchers and practitioners in related fields, enhance their climate literacy, data science proficiency and interdisciplinary communication skills. They also get to step outside the classroom setting to understand their impact.

One example of this holistic approach is a project team that focused on accurately mapping tree species at high resolution to evaluate biodiversity, helping inform forest management and conservation. The team developed an AI-driven deep learning framework that integrates hyperspectral imagery, aerial photography, LiDAR point clouds and ground data to classify tree species by analyzing their uppermost leaves and branches, or “crowns.” Compared with traditional approaches, this method allows for more precise identification even in complex, overlapping canopies.

Students trekked through Duke Forest to get a grasp of the overlapping canopies from the ground, gain insights from a Duke Forest expert and better understand how their data analysis would be working to support forest dynamics. By the summer’s end, students had contributed to a detailed map of a National Ecological Observatory Site, and during the fall semester, several students are continuing to build on this research.

“Supporting students on this project means giving them a unique opportunity to work at the intersection of ecology, AI and remote sensing—skills that are in high demand for the future of environmental science,” said project lead Tong Qiu, assistant professor of ecology at the Nicholas School of the Environment. “I hope their experience here becomes a springboard for their own research careers.”

Climate+ is offered by the Nicholas Institute for Energy, Environment & Sustainability in partnership with the Rhodes Information Initiative at Duke. The summer program is aligned with the Duke Climate Commitment, a university-wide initiative that unites Duke’s education, research, operations and public service missions to address climate challenges. Funding for Climate+ comes from The Duke Endowment and the Rhodes Information Initiative.