🌍 Inspiration

GeoGuessr, a geography game, challenges players to identify Google Street locations using visual clues, but they often rely on prior knowledge rather than deep learning insight. We noticed a lack of resources online to help beginner players, so we built GeoTrainer—an AI-powered geography game that not only tests players' knowledge but also explains why an image belongs to a specific country.

🔍 What it does

Players analyze Google Street View images and guess the country. Our AI provides feedback based on:

  • Road signs, languages, and text
  • Architecture and infrastructure
  • Vegetation, terrain, and climate
  • Street markings and traffic patterns

Unlike the existing GeoGuessr game, GeoTrainer teaches players how to improve over time through AI-driven explanations.

🛠️ How we built it

  • Node.js server handles requests and processes data.
  • Python Flask server sends images to OpenAI’s API for analysis.
  • OpenAI API extracts meaningful location-based hints from images.
  • Figma UI → React & CSS for a seamless and intuitive front-end experience.

🚧 Challenges we ran into

  • Matching GeoGuessr’s UI to replicate a familiar player experience.
  • Parsing AI-generated responses into relevant and understandable hints.
  • Making the game intuitive and accessible for both GeoGuessr veterans & new players.

🎉 Accomplishments that we're proud of

  • Effective UI/UX throughout the prototype.
  • Real-time AI feedback makes geographic learning more engaging.

🔮 What's next for GeoTrainer

  • Multiplayer challenges—compete with friends in realtime using WebSockets!
  • Multimedia hints like native language enunciation

Built With

Share this project:

Updates