OpenStreetMap Buenos Aires City Dataset Import

A toolset for importing, cleaning, and processing Buenos Aires building data from official government sources into OpenStreetMap-compatible format.

Context & Purpose

This project bridges official Buenos Aires building data with OpenStreetMap by:

• Using authoritative city dataset geometries and attributes
• Cross-referencing with live OSM data to prevent duplicates
• Cleaning and transforming data through SQL migrations
• Generating OSM changefiles chunked by city blocks ("manzanas")

Project Structure

• bin/ - Helper scripts:
  • download-dataset.sh - Fetches the Buenos Aires urban fabric dataset
  • import-datasets.sh - Imports city and OSM data into PostgreSQL/PostGIS
  • run-migrations.sql - Executes SQL migration scripts for data transformation
  • export-geojson.sh - Exports query results as GeoJSON for debugging
• examples/ - Example GeoJSON files for some of the addressed issues
• migrations/ - SQL and Python scripts for data transformation
• changesets/ - Output directory for generated OSM changefiles

Getting Started

1. Download Datasets

1. Go to Overpass Turbo

2. Paste the following query:

   [out:json][timeout:900];
   area["name"="Buenos Aires"]["boundary"="administrative"]["admin_level"="8"]->.searchArea;
   (
     way["building"](area.searchArea);
     relation["building"](area.searchArea);
   );
   out body;
   >;
   out skel qt;
   

3. Run the query, then use Export > GeoJSON to save as data/live-osm-data.geojson

4. Run the dataset download script:

bin/download-dataset.sh

This script will save the contents to data/tejido.geojson.

2. Import Datasets into PostgreSQL/PostGIS

Start the database:

docker compose up -d

Import the datasets:

bin/import-datasets.sh

3. Run Data Cleaning and Transformation

bin/run-migrations.sql

4. Generate OSM Changefiles

uv run migrations/500-create-changefiles.py

Changefiles will be available in the changesets/ directory, organized by city block.

Requirements

• Docker Compose for PostgreSQL/PostGIS
• Python 3.12+
• uv package manager
• GDAL with ogr2ogr

Visualization

To visualize the processing results:

bin/export-geojson.sh "SELECT * FROM processed_buildings LIMIT 100" > debug.geojson

Then open the file in QGIS or any GeoJSON viewer.

References

• Buenos Aires Urban Fabric Dataset

License

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

Contributing

Contributions are welcome! Feel free to:

• Report issues
• Suggest improvements to the processing pipeline
• Add new data cleaning rules
• Enhance documentation