UNIVERSITY PARK, Pa. — In December 2018, a team of five Penn State students from four separate colleges earned the top spot in the international Kaggle University Hackathon, the first-ever university hackathon of its kind.
Suraj Dalsania (information sciences and technology, College of Information Sciences and Technology), Neil Ashtekar (computer science, College of Engineering), Izzi Oakes (integrative arts, College of Arts and Architecture), Will Wright (mathematics, Eberly College of Science), and Ming-Ju Li (electrical engineering, College of Engineering) joined to compete for Penn State. They are all members of Nittany Data Labs, a College of IST student organization aimed at preparing students to become leading data scientists in a number of fields.
“I feel like NDL provides you with a foundation, or direction, in which to go to learn more things,” said Ashtekar, team lead. “It pointed us in the right direction to help us learn what we needed for this hackathon.
At the hackathon, teams were given a data set of medical drug reviews and ratings, indicating whether or not people found them helpful. Team NDL set off to address a set of problems, such as creating a way for people to find the best medication for their illness and determining which machine models work best for predicting a drug’s rating based on consumer reviews.
Their final submission was in the form of a code that used sentiment analysis to explore how it affected user ratings, the perceived usefulness of ratings, and which machine learning models could best predict the sentiment or rating based on the user review, among many other factors. To accomplish their task, the team members worked individually at first, but then combined their data together. They credit their interdisciplinary approach and the foundation they’ve built through classes and Nittany Data Labs with advancing the project.
“NDL got me exposed to a good foundation for how to think about Python, so I’ve taken advantage of that,” said Wright. “After the hackathon, I learned a lot more about different packages and what to look at more in-depth — working with something of that size, we had to work with larger packages [to code].”
“When it came to the predictive modeling part, we started working together and it was more organized,” added Oakes.
Teamwork is something that the team can credit to learning at Penn State, as that is one skill that has been taught through their courses and other experiences.
“In my time [at Penn State], I have developed beneficial skills in the art of networking and professional development” said Dalsania. “Being in the College of IST, as well as a part of NDL, has taught me real world lessons on how to interact with business professionals, how to work together on group projects, and how to become a better leader.”
In the hackathon, Team NDL scored highest out of the 18 participating teams among universities around the globe. They earned Kaggle swag and were featured on the Kaggle blog, No Free Hunch. Owned by Google LLC, Kaggle is an online community of more than 1 million data scientists and machine learners that offers a collaborative space and hosts competitions to solve data science challenges.