GeoGuessr tips with AI

This repository contains a growing set of tools aimed at understanding and boosting play on GeoGuessr, a game that challenges users to recognize locations from Google Street View's panoramic images.

The project was born out of a desire to help my kids better understand geography, but it has turned into an experiment in using AI tools to generate structured data — in this case, hints for identifying world countries and US states on the roadways.

The project, for now, includes scripts to collect and generate game-specific metadata about countries and US states. It also contains code to create a simple country mapping quiz and a filterable grid of national flags, with country-specific pages baked into static files.

It's very new and a work in progress, but live here: https://levelupgeo.com

Features

Ai-driven data generation

• Uses AI tools to create structured metadata for over 190 countries and all 50 US states.
• Generates clues for identifying locations on roadways, including details like language, signs, cars, and infrastructure.
• Employs natural language generation techniques to fill in geospatial details dynamically.

Front-end features baked with baker

• Leverages the LA Times' Baker tool for static site generation.
• Creates:
  • A filterable grid of countries with flag images.
  • Individual country pages featuring maps, metadata, and AI-generated clues.
  • A country mapping quiz for geography practice.

Duel analysis tools

• Scripts to fetch GeoGuessr duel details for deeper analysis.
• Compile your duels into a dataframe for statistical or visual exploration.

Geospatial integration

• Uses GeoJSON data to dynamically highlight countries and US states on interactive maps.
• Integrates Mapbox for visualizing country boundaries and adding inset regional maps.

How to use

Installation

1. Clone this repository:

   git clone https://github.com/yourusername/guessr-meta-generator.git
   

2. Install the dependencies:

   pip install -r requirements.txt
   

Running the project

Building the static site

1. Generate the entire site:

   python bake.py build_all
   

2. Serve the site locally for testing:

   python bake.py serve
   

3. Visit http://127.0.0.1:8000 in your browser to explore the generated pages.

Using the duel analysis tools

• Run the duel fetching script:

  python scripts/fetch_duels.py
  

• Compile a dataframe of all your duels:

  python scripts/compile_duels.py
  

Contributing

This project is a work in progress, and contributions are welcome! If you have ideas for new features or improvements, feel free to submit a pull request or open an issue.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Acknowledgments

• GeoGuessr for the inspiration behind this project.
• LA Times Baker for the static site generation tool.
• OpenAI for AI-driven metadata generation.