GreenCalculus open factor data

What the GreenCalculus corpus covers — 16,673 emission factors from 137 publishers, with each publisher's licence and whether it may be republished.

Every row is generated from the live API by build.py. Nothing is hand-maintained: the licence filter is each publisher's own terms as recorded in the API response, so an upstream licence re-review changes this export on the next run.

Data version 2026.188 (2026-09-10).

What's here

coverage/index.csv — 16,673 rows, one per factor key. Key, name, section, unit, gas, GHG Protocol scope, publisher, licence, whether it's redistributable, and why not when it isn't.

No factor values. This is GreenCalculus's own metadata about what the corpus covers, not any publisher's data — so it is publishable regardless of upstream terms, and you can use it freely to answer "is there a factor for X?" without worrying about anyone's licence.

LICENCES.md — every source, its publisher, its licence, and the attribution line that licence requires. Plus every excluded source with the publisher's own stated reason, verbatim.

Want the numbers?

The values are not in this repo. Look any factor up from the API — no API key needed, the corpus is open to read:

curl "https://api.greencalculus.com/v1/factors?key_prefix=grid.gbr.electricity.location_based&limit=1"

pip install greencalculus

from greencalculus import GreenCalculus
f = GreenCalculus().factor("grid.gbr.electricity.location_based")   # no key
print(f["value"], f["unit"])                       # 0.13096 kg CO2e per kWh
print(f["factor"]["source"]["cell_ref"])           # 'UK electricity'!E25
print(f["factor"]["citation"]["proof_url"])        # a page your reader can check

Every factor also has a permanent verification page showing the publisher, the document, the exact cell and whether it may be republished — verify.greencalculus.com/<key>.

The split

Of 16,673 factors across 137 sources:

• 15,347 factors / 75 sources carry a licence that permits republication (Open Government Licence, CC BY 4.0, Etalab, US public domain, Eurostat reuse…).
• 1,326 factors / 62 sources do not, and the reason is the publisher's own: paid standards (ISO, EN 15978), NonCommercial or NoDerivatives terms (PCAF, SBTi, WBCSD Pathfinder), share-alike that would be viral over your results (EXIOBASE, ecoinvent), or simply no grant found (ICAO).

Two sources are held back by hand, for different reasons:

• IEA_AI_ENERGY_2025 is flagged CC BY 4.0 upstream, which conflicts with our reading of the IEA's terms on third-party footprints. It stays out until that is resolved with the publisher.
• LOVEHAGEN_2023_EMBODIED_USER_DEVICES is CC BY-NC-ND 4.0, and the paper's first author granted GreenCalculus written permission to display these seven values and to serve them through our API. That permission runs to us; it is not ours to pass on to you, so the rows are served and citable but not offered for republication.

The second is the general case worth knowing about: the redistributable flag answers may GreenCalculus serve this row, which is not always the same question as may you republish this value. Where the two diverge, the DENY map in build.py records the divergence and its reason. It can only ever remove a source, never add one, so a mistake there costs coverage rather than causing a licence breach.

Rebuild

python3 build.py                # coverage/ + LICENCES.md   (what this repo ships)
python3 build.py --with-values  # also open/ — the republishable subset WITH values

open/ is gitignored. Whether to publish values is a deliberate decision, not a build flag away from happening by accident.

Licence

Code: MIT. coverage/index.csv is GreenCalculus metadata, published CC BY 4.0 — attribute GreenCalculus (greencalculus.com). Factor values obtained from the API carry the licence of their underlying source, named in every response.