Hometown Success Engine

Discover how America's geography could help find the next generation of Team USA Olympic and Paralympic talent.

Comment

Challenge 2

🏅 Inspiration

As a fan of the Olympics and Paralympics, I always wondered why certain states seem to produce so many Team USA athletes. Is it the climate? The geography? The culture? I wanted to build something that could help fans explore that question in a data-driven, visual way — and the Team USA x Google Cloud Hackathon was the perfect opportunity.

🗺️ What It Does

The Hometown Success Engine is an interactive web app that lets fans explore which U.S. states could be considered "hubs" for Team USA Olympic and Paralympic talent. Users can filter by sport, region, and athlete type (Olympic, Paralympic, or both) to see how geography and climate could help find talent pipelines across America.

The hub score for each state is calculated as:

$$\text{Hub Score} = \frac{\text{Total Team USA Athletes}}{\text{State Population}} \times 100{,}000$$

Key features:

  • 🏆 State-by-state hub rankings based on total Team USA athletes per capita
  • 🏊 Sport breakdowns showing which disciplines each state excels in
  • 🤖 Gemini-powered insights that generate conditional, data-driven narratives for each state
  • ♿ Paralympic parity — Paralympians are tracked and displayed with equal prominence alongside Olympians
  • 🌦️ Climate and geography overlays correlating public weather data with sport clusters

🛠️ How I Built It

Layer Technology
Frontend Next.js + Tailwind CSS
AI Gemini API via Vertex AI
Data teamusa.com + open source Olympic/Paralympic repos
Geography/Climate NOAA + U.S. Census public data
Deployment Google Cloud Run

⚡ Challenges

  • Ensuring Paralympic parity in the data — Paralympic athlete records required extra effort to source and normalize alongside Olympic data
  • Writing Gemini prompts that consistently used conditional language ("could help find," "may support") without implying geography guarantees results
  • Staying strictly within data use restrictions — no finish times, no individual NIL, US-scope only

📚 What I Learned

  • How to use the Gemini API for narrative generation in a real production app
  • How to deploy a Next.js app on Google Cloud Run
  • How to think about data storytelling responsibly — especially around athlete representation and avoiding misleading conclusions

Built With

  • gemini-api
  • google-cloud-run
  • next.js
  • noaa-api
  • python
  • tailwind-css
  • typescript
  • u.s.-census-api
  • vertex-ai
Share this project:

Updates