In 2022, four students in the inaugural cohort of the Bren School's Master of Environmental Data Science program set out to answer a deceptively simple question: what determines where utility-scale wind and solar energy projects get built in the United States?
The project, advised by Bren Professor Ranjit Deshmukh and developed in partnership with UCSB's Environmental Studies Professor Grace Wu, sought to identify the factors that influence where utility-scale onshore wind and solar projects are built across the United States. Through collecting and analyzing spatial data, the team — Paloma Cartwright, Joe DeCesaro, Daniel Kerstan, and Desik Somasundaram — examined how a range of environmental, economic, and infrastructure factors relate to existing project locations. They used those insights to develop statistical and machine learning models that predict where future wind and solar development is most likely to occur and created an interactive dashboard to help users explore the results.
The team designed the project with future use in mind. "The plan was always to have our work used in a publication, and we intentionally wrote the code for the analysis to be easy to rerun with new layers and on different machines," DeCesaro said.
Their work went on to serve as the foundation for a peer-reviewed study published April 28, 2026, in the journal Environmental Research Letters, titled "Factors Shaping the Siting of Utility-Scale Solar and Wind Projects in the United States."
The published study was led by UCSB Environmental Studies Professor Grace Wu and co-authored by Deshmukh, the four alumni, MESM student Henry Strecker, and Yohan Min of UCSB's Environmental Studies Program. Building on the capstone foundation, the team used an ensemble of machine learning and statistical approaches to analyze projects built between 2017 and 2023. The findings reveal a meaningful contrast between the two technologies. Solar siting is more flexible, shaped by a wider range of factors including proximity to population centers, road accessibility, and areas with lower ecological sensitivity. Wind siting, by contrast, is primarily driven by wind resource quality and agricultural land use — a narrower set of constraints that concentrates development in fewer regions. For both technologies, proximity to existing transmission infrastructure was among the most consistent predictors of where projects land.
The study also examined equity dimensions of siting. For solar, the research found that federally designated disadvantaged communities in the Midwest, South, and Northeast were significantly less likely to host new projects than their non-disadvantaged neighbors, with the most dramatic gap in the Northeast. For wind, the pattern was most consistent in the Midwest, where higher siting probabilities in non-disadvantaged communities held across all modeling approaches. While reduced development pressure may limit land-use burdens on disadvantaged communities, it could also mean fewer opportunities to benefit from the jobs, tax revenues, and other economic gains associated with renewable energy development.
For DeCesaro, the publication is a reflection of both the team and the program. "Working with this group of people made it all possible and our client and advisors were extremely helpful and insightful," he said. "We had a clear idea of what we wanted to execute from the beginning which made it much easier to plan for and develop. Having the work published in a scientific journal was gratifying and is a testament to our own hard work and how the program prepared us to execute the needs of our client."
As renewable energy deployment accelerates across the U.S., the study offers insights into the factors shaping where new projects are built and provides tools that can help inform future planning decisions. The paper is open access, and the team's analysis code, national siting prediction maps, and interactive dashboard are publicly available.
