Skip to content

Coupling with OpenStreetMap (OSM)

One of the powerful features of DamageScanner is the ability to work with OpenStreetMap (OSM) data as exposure input. This enables fast and global-scale exposure mapping, using freely available, crowdsourced infrastructure data.

This page outlines how to prepare and use OSM data for damage assessment workflows, using .osm.pbf files directly.

Why Use OSM Data?

  • Freely available, global coverage
  • Includes many infrastructure types (roads, buildings, energy, healthcare, etc.)
  • Constantly updated by the global community
  • Works well with vector-based workflows in DamageScanner

⚠️ Note: OSM data quality and completeness can vary by location. Always validate your input.

How DamageScanner Works with OSM

DamageScanner includes built-in support for working directly with .osm.pbf files. The tool provides utilities to:

  • Download .osm.pbf files from Geofabrik
  • Extract specific infrastructure classes
  • Clean and preprocess the data into usable GeoDataFrame or .gpkg

These functions are implemented in: - osm.py - download.py

✅ This approach allows for highly automated and reproducible integration of OSM infrastructure.

You can extract features like this:

from damagescanner.osm import read_osm_data

# Read critical infrastructure data from an OSM .pbf file
osm_path = "data/netherlands-latest.osm.pbf"
features = read_osm_data(osm_path, asset_type="road")

The asset_type must be one of the predefined keys such as road, rail, power, healthcare, education, etc. The output is a cleaned GeoDataFrame with valid geometries and an object_type column.

Manual Tagging (Optional)

If you're preparing custom OSM files manually (e.g. from .shp or .gpkg):

# Reproject and tag manually loaded GeoDataFrame
gdf = gdf.to_crs("EPSG:32633")
gdf = gdf[gdf.geom_type.isin(["LineString", "MultiLineString"])]
gdf["object_type"] = "road"

⚠️ The required column name is always object_type — this is how DamageScanner matches features with vulnerability curves and max damage values.

Save to file if needed:

gdf.to_file("data/cleaned_osm_roads.gpkg", driver="GPKG")

Example Usage with DamageScanner

from damagescanner import DamageScanner

hazard = "data/flood_depth_100yr.tif"
osm_exposure = "data/cleaned_osm_roads.gpkg"
curves = "data/vulnerability_curves.csv"
maxdam = "data/maxdam.csv"

scanner = DamageScanner(hazard, osm_exposure, curves, maxdam)
damage = scanner.calculate()

This calculates direct damages using cleaned vector infrastructure data from OSM.

Tips for Reproducibility

  • Document the OSM extract date and source (e.g. Geofabrik region)
  • Use consistent object naming (e.g. road, bridge, hospital) to match your vulnerability inputs
  • Save your processed data in .gpkg or .shp format for reuse

📚 See Also