Winners 🔥

After 3.5 hours of team presentations and nearly an hour of thoughtful deliberation by the judges, 4 standout teams (out of 13) emerged as the winners of this year’s DataFest. Teams were evaluated on the quality of their recommendations, soundness of analysis, clarity of presentation, data visualization, use of external data, and overall creativity.

1. First Place Award - GGAP !!!!

(Alyssa, Paisley, Grace & Grace)

The 2025 epilogue to the GGAP 303 sequence trilogy

Mentor: Cathy Kim

GGAP stood out for their creativity, clarity and polish!—earning the top spot on four of five judges’ lists. Not only did they build a product ready for real-world use, but their explanation of the technical details was both clear and compelling. And just when the judges thought it couldn’t get better, GGAP brought the house down with a clever roleplay—two teammates acted as clients, while the others demoed the product like it was a live pitch. It felt like a real client meeting!

Technical aspect: Used \(k\)-means clustering to identify leases similar to client requirements and provide recommendations. The clusters and the recommended leases were visualized using a dashboard. A couple of demos were presented, where the client inputs the requirements in the dashboard, and then the algorithm visualizes the results and the recommendations.

2. Second Place Award - Team NA !!!

(Noah & Aryaman)

Filling in the gaps, one data point at a time

Mentor: Zihan Zhao

The judges were deeply impressed by the analytical rigor demonstrated by NA! Unlike many teams who chose models more instinctively, NA took a methodical approach—testing several models and combining them into a robust ensemble. What further set them apart was their thoughtful integration of highly relevant external data, which significantly enriched their analysis.

Technical aspect: Developed a times series model to forecast the volume of leases using quarterly market characteristics. The model was an ensemble of KNN, Poisson, ElasticNet, XGBoost, and Random Forest. Recommendations were based on market trends, and important features effecting lease volume of the specific industry.

3. Third Place Award - Hot Tea !!

(Ice, Robert, and Owen)

Just an ordinary dataset, but with the right model, we can make it trend like a hot tea

Mentor: David Frost

Hot Tea impressed the judges with the originality of their analytical approach. They crafted a truly unique question that offered a fresh perspective on the open-ended problem, making their analysis stand out as both memorable and impactful.

Technical aspect: Focused the analysis on the impact of wildfires on California leases, visualized their effect on leasing trends, and provided recommendations on optimal timing, location, and lease types based on wildfire occurrences. This solution shows that simple analysis can also lead to useful insights.

4, People’s Choice Award - The Prompt Engineers !

(Harrison, Abigail, Jinhe, and Olzhasbek)

We may not have the answers, but we will be sure to phrase a question well enough to fake it

Mentor: Xinhui Qian

The Prompt Engineers stole the show right from the start with a rap-style team intro and wrapped things up with a custom song for the client—winning over the crowd and clinching the award! But it wasn’t just the music that stood out. While the award was audience-driven, the judges were equally impressed by the team’s sharp local insights, especially when one member used their on-the-ground knowledge to deliver a spot-on answer during Q&A!

Technical aspect: Visualized leasing trends before and after COVID, which motivated the team to focus on cost-effective space utilization. Broad-level recommendations were provided based on the market trends. This is another solution showing that simple analysis can lead to useful insights.