The 2026 Wharton High School Data Science Competition, presented by Google Gemini brought together more than 700 teams from 48 countries, each tackling the challenge of analyzing simulated ice hockey data from the “World Hockey League” to uncover meaningful insights about the sport. Across submissions, students built models to evaluate team strength, predict game outcomes, assess offensive line quality, and visualize complex relationships among key measures. The strongest teams went beyond technical accuracy. They communicated their ideas with clarity and confidence, grounding their decisions in both hockey knowledge and thoughtful reasoning. A central takeaway from this year’s competition was clear: in real-world data science, there is rarely one “right” answer. That ambiguity can be challenging, but it is also what makes the field so creative, dynamic, and impactful.

Our team of expert data scientists benchmarked submissions and was impressed by the overall quality and sophistication of the work. Top-performing teams consistently demonstrated careful data cleaning, creative and intentional feature engineering, and disciplined use of statistical methods. They also showed a strong awareness of the nuances that shape the game, incorporating factors such as goaltending, empty-net situations, power plays, opposing team strength, and time on ice. The very best teams struck a thoughtful balance between simplicity and strategy, finding elegant ways to capture the complexity of ice hockey while delivering clear, compelling insights.