On the weekend of October 4-6, 2024, the Harvard Data Science Initiative (HDSI) hosted its first-ever datathon in collaboration with the U.S. Department of Agriculture’s National Agricultural Statistics Service (USDA NASS). Through the weekend, 20 teams of undergraduate and graduate students from around the Greater Boston Area tackled pressing data science challenges at the intersection of agriculture, food security, and climate science.
“A datathon is like a hackathon, but more focused on challenges that arise when working with data,” explains Jennifer Chow, HDSI Director of External Engagement and Agri Datathon organizer. “Real world data is imperfect, it’s messy – you grapple with it a bit before the solutions reveal themselves. Working with it is a really different experience compared to being in a classroom and submitting a problem set.”
Agriculture-Focused Challenges
Technical challenges were designed by Elaine Swanson, a computational engineer who specializes in agriculture research and recent graduate of the Masters in Computational Engineering Program from the Harvard John A. Paulson School of Engineering and Applied Sciences. “I was inspired by the USDA’s 2022 Agricultural Census Release session I attended during the 2023 Agricultural Outlook Forum (AOF).,” Swanson shares. “At AOF, I also learned about the USDA’s initiative called NIFA’s NextGen, which engages with higher education institutions, but primarily Land-Grant Universities. Harvard, MIT, and Boston-based universities are, of course, not among them. After going through Harvard’s program, I recognized the state-of-the-art skills students were developing, but a lack of focus on agriculture climatology research. So I saw this as a unique opportunity to introduce an important subject matter, and the wealth of USDA’s data, to students in universities that have less engagement with the USDA.'”
Collaborating with Ginger Harris of USDA NASS, Swanson devised three prompts that reflect real-life challenges that agricultural data scientists might encounter:
- Geospatial: Developing frameworks that integrate satellite imagery datasets to analyze crop phenology and predict NDVI trends.
- Demographics: Exploring the relationship between farmer age and agricultural production across the contiguous U.S. to understand regional variations and economic impacts.
- Crop Diversification: Investigating the historical and spatial relationship between crop diversification, land value, and production across the U.S.
Teams had 60 hours to build and present solutions. “The submissions we received demonstrated a wide range of depth and methods.” Swanson says. “I enjoyed comparing my own initial approach to how the teams solved the problem. Some solutions were more streamlined, scalable, and achieved better accuracy. I knew from the start that I would be learning from the process as well.”
Bringing Together a National Network of Experts
To mentor teams and judge their outputs, the HDSI recruited a team of in-person and virtual data experts across fields such as agriculture and agronomy, climate science and environmental studies, geographic information systems, and AgTech including precision agriculture. Chow says, “We were overcome by everyone’s generosity with their time and expertise. Given that HDSI is supporting work at the critical intersection of climate, agriculture, and food security, I hope this is just the first chance to bring folks together.”
Mentors
- A.J. Kumar, IndigoAg
- Dan Sumner, UC Davis
- Evan Marshall, Massachusetts Dept. of Agricultural Resources
- Hanbin Lee, UC Davis
- Jeffrey Hunt, USDA
- Johannes Knittel, Harvard
- Karen Olsen, CIBO Technologies
- Margaret Kosmala, CIBO Technologies
- Michelle Audirac, Harvard
- Mauricio Tec, Harvard
- Oladimeji Mudele, Harvard
- Shane Bussmann, CIBO Technologies
- Siqin (Sisi) Wang, USC and Harvard
- Virginia Harris, USDA
- Wolfram Schlenker, Harvard
Judges
- Andrew Dau, USDA
- Angela J. Rigden, UC Irvine
- Audra Crawford, USDA
- Bryan Combs, USDA
- Cathy Huyghe, Enolytics
- Charles Taylor, Harvard
- Claudio Battiloro, Harvard
- Denise Abreu, USDA
- Jake Lehan, Massachusetts Department of Agriculture
- John Quackenbush, Harvard
- Joseph Dexter, Harvard
- Lu Chen, USDA
- Luca Sartore, USDA
- Peter Huybers, Harvard
- Rob Paarlberg, Wellesley College
- Slide Kelly, Harvard
- Valbona Bejleri, USDA
Advisory Committee
- Christopher Golden, Harvard
- Colette Basiliere, Public Interest Technology New England
- Rick Mueller, USDA
- Stratos Idreos, Harvard
Organizing Committee
- Elaine Swanson, Harvard
- Ginger Harris, USDA
- Jennifer Chow, Harvard
- Torey Lawrence, USDA
Announcing the HDSI Agri Datathon Winners
Congratulations to the winners! Winning projects were chosen by the panel of judges based on innovative problem-solving, data visualization, and storytelling through data. Explore their submissions below:
First Place – CodeCultivators ($900 Prize)
Team Members: Krishna Venkatesh, Laasya Anantha Prasad, Rakshak Kunchum
Analyzing Demographic Shifts and Their Economic Impact on U.S.Agriculture
Second Place – Rug Pullers ($600 Prize)
Team Members: Ryan Gilbert, Reetom Gangopadhyay, Shreyas Puducheri
NDVI Forecasting for Agricultural Crops
Video coming soon
Third Place – NU Field Analysts ($300 Prize)
Team Members: Atyab Hakeem, Nick Cantalupa, Sean Duffy
Midwest Agriculture: Crop Patterns, Land Use, Revenue, and Climate Trends
Honorable Mention, Best Presentation – Feed the Future
Team Members: Anmar Abdi, Oluwaseun Ademiloye, Vijay Fisch
Navigating Urbanization: The Impact on Crop Diversity and Agricultural Productivity in Indianapolis
Honorable Mention, Best Visualization – Huskies
Team Members: Aditya Singh, Khushi Neema, Sarthak Khandelwal
Age Demographic Trends in U.S. Agriculture: A Spatiotemporal Analysis of USDA Data