osm-powerplants 0.1.5

Extract and process power plant data from OpenStreetMap

OSM Power Plants

Extract power plant data from OpenStreetMap for energy system modeling.

Installation

pip install osm-powerplants

Quick Start

Command Line

osm-powerplants process Germany France -o europe.csv

Python API

from osm_powerplants import process_units, get_config, get_cache_dir

df = process_units(
    countries=["Chile", "Greece"],
    config=get_config(),
    cache_dir=str(get_cache_dir(get_config())),
)

Why some plants are excluded

OSM is a crowd-sourced dataset, and many power=plant elements lack the metadata needed to plug them into an energy system model. Every OSM element that cannot be resolved into a usable unit is dropped from the output and recorded in a rejection report. The dominant reasons are:

Reason	What it means
Missing output tag	Plant has power=plant but no plant:output:electricity (and no source-specific capacity-estimation default in config.yaml). Very common on solar and geothermal farms.
Capacity placeholder value	plant:output:electricity is set to a stub like yes instead of a number.
Capacity regex no match / Capacity non-numeric	Capacity tag exists but cannot be parsed (unusual units, free-form text).
Missing source tag / Missing technology tag	Cannot classify the plant — these can be relaxed with missing_technology_allowed: True in config.
Element within existing plant geometry	A generator polygon lies inside an already-processed plant boundary (deduplication).

If a country returns fewer plants than you expected from overpass-turbo, set rejected_output_path to see exactly what was dropped and why:

config = get_config()
config["force_refresh"] = True  # rejection data is only populated on API fetches

df = process_units(
    countries=["Kenya"],
    config=config,
    cache_dir=str(get_cache_dir(config)),
    output_path="kenya.csv",
    rejected_output_path="kenya_rejected.csv",  # writes CSV + .geojson
)

The CSV lists each rejection with the OSM URL, coordinates, and reason code; the sibling GeoJSON can be loaded in JOSM or QGIS so OSM contributors can triage the gaps directly.

Output Format

Column	Description
projectID	OSM-based identifier
Name	Plant name
Country	Country name
lat, lon	Coordinates
Fueltype	Solar, Wind, Hydro, Nuclear, Natural Gas, etc.
Technology	PV, Onshore, Run-Of-River, Steam Turbine, etc.
Set	PP (power plant), Store (storage)
Capacity	MW
DateIn	Commissioning year

Documentation

Full documentation: https://open-energy-transition.github.io/osm-powerplants

Global dataset

A ready-to-use global dataset, regenerated monthly by CI, lives at the repo root:

• osm_global.csv — merged plants for all 16 regions (~250 countries).
• osm_global_rejected_plants.csv — features dropped by quality filters, with reason and coordinates.
• osm_global_rejected_plants.geojson — the same rejections as mappable points for OSM-contributor triage.

Per-region outputs (one triplet per region) live under datasets/. Region definitions are in src/osm_powerplants/regions.py.

The CI workflow (.github/workflows/data.yml) runs on the 1st of each month: one parallel job per region (uploads artifacts) followed by a merge job that commits the global outputs.

Local regeneration:

python scripts/extract_region.py --region europe --clear-cache
python scripts/merge_global.py

Integration with powerplantmatching

This package provides OSM data for powerplantmatching. osm_global.csv is consumed by PPM's matching pipeline. The evaluation that selected which countries should include OSM by default lives in evaluation/ — see evaluation/osm_global_report.md for the methodology and findings.

The legacy osm_europe.csv at the repo root is frozen: it is retained only while the current PPM config still points to it and will be removed after the PPM PR that re-points to osm_global.csv merges.

Development

git clone https://github.com/open-energy-transition/osm-powerplants.git
cd osm-powerplants
pip install -e ".[dev]"
pre-commit install
pytest

Acknowledgments

This project is developed and maintained by Open Energy Transition, a company dedicated to accelerating the global energy transition through open-source tools and data.

License

MIT License - see LICENSE for details.

Data sourced from OpenStreetMap © OpenStreetMap contributors.