Inspiration
I recently moved and my new home office came with a soccer ball patterned light switch! So when I saw that one of the curated datasets for the Hex-a-thon was about soccer, I knew it was destiny. As for the design, I love "data art" and after playing with Hex and seeing what beautiful visualizations I can build with it, I wanted to make something inspired by that.
What it does
This analysis addresses two interconnected ideas: Analyzing how soccer has evolved since its codification in 1863, and predicting which team will win the 2026 FIFA World Cup - the first to feature 48 teams across 12 groups.
How I built it
I used Hex magic and the Notebook Agent significantly to help me quickly build out the 2026 FIFA prediction, but outside that, I wanted to see what I could build in Hex all on my own, and was very satisfied with what I was able to achieve.
Approach & Methodology: For my analysis on soccer evolution, I matched up changepoints in the game data provided with major events in world history. I also analyzed all historical match data (1872-2025) to identify long-term trends in scoring patterns, defensive evolution, and competitive balance. For the 2026 prediction, I developed a Monte Carlo simulation framework with 10,000 tournament runs, incorporating opponent-adjusted team ratings that account for strength of schedule, traditional World Cup pedigree, and recent performance (2020-2025). This methodology corrects for a critical flaw in some naive approaches I found early in which my model was inflating statistics for teams who play against weak opponents.
Challenges I ran into
The earliest version of my predictive model predicted that New Zealand would win the FIFA world cup. I knew that this couldn't possibly be correct, so I asked the notebook agent to explain how it got those results. I couldn't believe how capable the notebook agent was at analyzing where it went wrong and adjusting the method to account for NZ's history against weak teams. I could speak to the Agent like I would another analyst and it would correct itself in real time.
Accomplishments that I'm proud of
I really played around with many different ways to visualize data in Hex - utilizing buttons as "tabs" for the same graph, and trying several different complex graph styles, really remixing on how the data was portrayed. I especially love the timeline of soccer viz I created, and ended up using it as the cover page for the report.
What I learned
Historical Evolution Findings: High-scoring games (7+ goals) declined from 10-15% in early soccer to 2-5% today, reflecting greater defensive organization. The percentage of scoreless games has gradually increased since the 1940s, now hovering around 10%. Modern soccer exhibits more competitive balance, with win percentages stabilizing at 77-80%. I also found some key players from soccer history that I previously had not heard of - like Ali Daei - and I am glad I had the opportunity to highlight them.
FIFA 2026 Prediction: I predicted that Spain would win the 2026 FIFA World Cup. Spain benefits from the highest adjusted rating and a favorable Group C draw. Outside of this, I also found that European dominance is overwhelming - UEFA teams command over 75% of total win probability, with England and Netherlands rounding out the top three. Traditional powerhouses Brazil and Argentina face surprising challenges at 2% and 3% respectively, while host nations (USA, Mexico, Canada) combine for less than 1%. Finally, the top five teams control over 75% of all win probability, suggesting a relatively predictable tournament.
What's next for Evolution of Soccer
I know for a fact that I will be borrowing some of the visualization styles I created for this report and using them in my day-to-day work as a data analyst. I can't wait to see what's next for the future of Hex! And you bet I will be tuning in to FIFA 2026 to see if my prediction was correct.
Log in or sign up for Devpost to join the conversation.