Independent Research Course Student Contributes to Large National Study of Beaver Dams
Most attempts to restore damaged ecosystems rely on human intervention, but North America’s most effective environmental engineers might be its beavers.
When CDS undergraduate student Chris Kong first joined CDS Assistant Professor of Psychology and Data Science Grace Lindsay’s research project through CDS’ new Independent Research course, he didn’t expect to spend the next year building tools to track beaver dams across the United States. “At that point, I didn’t expect to extend this research to this long,” Kong said. “But after that semester, we believed the project could carry on even further.”
The project originated with the non-profit Collaborative Earth, where Lindsay works with a team of volunteer researchers studying beaver dam impacts. “Beavers re-engineer waterways in ways that can impact vegetation, soil properties, flooding, wildfire, and more,” Lindsay said. “Land managers are using them as part of nature-based restoration programs. But to use beavers most effectively, we need to verify the effects of past restoration programs and predict the impacts of future ones.”
The project applies computer vision machine learning methods to satellite and aerial imagery data to locate beaver dams across the United States and understand how these structures affect their surroundings. Through analyzing this data, the team identified patterns in how beaver-engineered changes to the landscape appear from above, and how these effects differ across various geographic regions.
To accomplish this, Kong helped develop a specialized pipeline for processing Landsat satellite data, particularly measures of vegetation. “Through the data he is collecting, we can compare vegetation metrics across locations on waterways with and without beaver dams,” Lindsay said. The team plans to share their code so others can conduct their own analyses.
Early findings suggest areas with beaver dams show increased vegetation coverage during summer months. “There are currently no large-scale datasets about beaver dams,” Kong said. “We want to build a dataset to be accessible and easily understandable to public sectors and other stakeholders of land.”
The connection between Kong and Lindsay’s research came through a new course at CDS. CDS Clinical Assistant Professor & Director of Undergraduate Studies Louis Mittel’s Independent Research course was created as an upper-level elective for advanced data science majors. The course matches talented undergraduates with faculty research projects across NYU, from CDS to sociology, economics, and the medical school, assigning each student their own individual project.
“The most impactful part of the classroom experience was the supportive and collaborative atmosphere within our cohort,” Kong said. “Since we all knew each other and frequently shared our projects and research experiences, I felt a strong sense of community. This was especially valuable as an undergraduate researcher, where confidence in understanding what research entails and how to overcome its challenges can often be low.”
While students work independently on their projects, they meet regularly as a group to discuss challenges and share solutions. “When students encounter challenges, we discuss those challenges, and students get feedback from their peers,” Mittel said. “They learn more about other people’s techniques, and recognize that other people’s problems are also their problems.”
The course structure also adapts to students’ needs. When Mittel notices common challenges, like handling missing data or creating effective visualizations, he designs lectures to address those specific issues. CDS PhD student Daniela Pinto Veizaga served as teaching assistant, providing technical guidance and mentorship.
Kong was the only junior in that first cohort of about 13 students, which gave him the opportunity to continue the research beyond the semester. Two other NYU undergraduates, Frederic Dai and Joyce Zhang, also joined the project after taking Lindsay’s Machine Learning for Climate Change course. The team continues gathering data, with plans to conduct broader analyses using machine learning tools in coming months.
“If we prove they can prevent wildfires and restore ecosystems at scale, policymakers can take action,” Kong said. “Maybe they can prevent beavers from being hunted or killed.”
For Kong, the course provided more than just research experience. “I attended Daniela’s recitation sessions each week, and we would often discuss topics after class,” he said. “She provided both technical guidance on my research and valuable advice on the graduate application process.”
The success of Kong’s project demonstrates the course’s ability to create meaningful connections between undergraduate talent and faculty research. As more students move through the program, they’re not just advancing their own careers — they’re contributing to research that makes a difference in the real world.
By Stephen Thomas
