Skip to main content

SDSU student team wins "Data Hackathon"

Hackathon
Jack's Hack, pictured above, recently won a Data Hackathon competition hosted by the University of Nebraska-Lincoln. Team members include (from left to right) Andrew Engel, Skye Brugler, Dwairka Bhattarai and Shailesh Pandit.

Jack’s Hack, a South Dakota State University student team led by Dwarika Bhattarai with Shailesh Pandit, Skye Brugler and Andrew Engel as team members, recently won the University of Nebraska-Lincoln’s Data Hackathon. The team won a $400 check and a travel trophy with their first-place prize, defeating teams from UNL, Kansas State University and the University of Illinois. 

“It was fun,” Pandit said. “It was a great learning opportunity for me and the other team members.”

The competition, which was sponsored by Corteva Agriscience, was part of the Nitrogen Use Efficiency workshop hosted by UNL’s Department of Agronomy and Horticulture on Aug. 1-3.

Each team was given a large data set in which they were to clean, analyze and distill information. Using only the team members, they were required to develop a new nitrogen fertilizer recommendation tool using the provided data—in only 15 days. 

The data sets were from 49 different sites and included weather history, corn yields, site management and soil properties. 

“Using that data set, we were asked to develop a model so that we can predict the economic optimum nitrogen rate for the corn in an unknown field,” Bhattarai explained.

Whichever team created a model that was the most accurate would win the competition. The tool would be evaluated based on the use of soil and crop science, accuracy and commercial feasibility. 

As Pandit explains, the team deployed a “simple plan” which allowed them to successfully attack the mountain of data. 

“We decided we were going to use a simple plan and that simple plan worked,” Pandit said. “The other teams had a complex idea, but our simple plan is how we won the competition.” 

Part of the plan included modeling the data and incorporating machine learning steps.

“We focused on the objective of the competition and tried to fit various models including stepwise regression, logistic regression and random forest until we got the best one rather than using a single complex algorithm,” Bhattarai added.

Pandit said the project was particularly helpful for him because the work done will prepare him for future projects while he pursues a master’s degree at SDSU.

“Going to that competition has given me a lot of ideas for my future academic plans,” Pandit said. “It was very motivating.”  

Oklahoma State University is scheduled to host next year’s workshop and competition. 

Bhattarai is looking to form another group that focuses on discussing and solving different statistics and data questions related to the student’s research. Anyone interested in being a part of this group can contact Dwarika Bhattarai via his email Dwarika.bhattarai@sdstate.edu.