fAIr Documentation

Appearance

User Manual for fAIr ​

Welcome! This manual is your step-by-step guide to getting started with fAIr, the AI-assisted mapping platform from the Humanitarian OpenStreetMap Team (HOT).

Table of Contents ​

Prerequisites ​

  • Stable internet connection
  • Basic mapping knowledge (learn more)
  • Basic understanding of AI datasets and models
  • Account on fAIr

Video Tutorials ​

Watch our step-by-step video guides:

Creating a Project in fAIr ​

  1. Log in to fAIr.
  2. Click Start Creating Dataset.
  3. Click Create New to make a new dataset.
  4. Name your dataset and click Create Training Dataset.
  5. Open OpenAerialMap in a new tab, find a suitable drone image, and copy its TMS URL.
  6. Paste the TMS URL into the Open Aerial Imagery tab in fAIr.
  7. Zoom to your image layer for visualization.
  8. Use the map tools to draw your Area of Interest (AOI).
  9. Click Fetch OSM Data to import existing building footprints for your AOI.
  10. Zoom in and review the imported buildings for accuracy.
  11. Edit and align building labels as needed for best training results.
  12. Upload your label fixes to OSM.
  13. Back in fAIr, fetch updated OSM labels for your AOI.
  14. On your dataset page, click View Models and then Create New to start a new model.
  15. Fill in model metadata and click Create AI Model.
  16. On the model page, submit a new training job with your chosen parameters.
  17. Monitor training progress and review accuracy graphs.
  18. Once satisfied, click Publish Training to make your model available for mapping.

Starting Mapping in Your Project ​

  1. On the model page, click Start Mapping next to your published training.
  2. Zoom to your area of interest and click Detect to run predictions.
  3. Review the AI-generated predictions on the map.
  4. Edit, accept, or discard predictions as needed.
  5. Push your validated results to OSM or export for further editing.

Editing Model Predictions ​

  1. Select a model and click Start Mapping to view predictions.
  2. Zoom in and review the predicted features.
  3. Accept or discard parts of the prediction, then re-run if needed.
  4. After making changes, click Submit my Feedback.
  5. All feedback is reviewed and validated by the project manager.

Note: Each "task" refers to a section of the map outlined with dotted lines and a number tag.

Help and Support ​

If you need help:

Thank You ​

Thank you for joining the fAIr community!
Your mapping efforts help create detailed, up-to-date maps that support crisis response, infrastructure planning, and community resilience worldwide.

By mapping with fAIr, you’re part of a global network of volunteers using open data and AI to make a positive impact.
Explore projects, join tasks, and connect with fellow mappers. Your contributions are invaluable!

Happy mapping!

The fAIr Team