mapbox/gabbar

Gabbar is an experimental machine learning tool developed by Mapbox to detect potentially harmful or suspicious edits in OpenStreetMap by classifying changesets as problematic or not. It is designed for OpenStreetMap contributors, validators, and maintainers who want to safeguard the integrity of the map by identifying anomalous edits. The tool provides a pre-trained binary classifier and can be used via a command-line interface to analyze specific changesets.