🔨 Derive nuclear status from EIA instead of hardcoded lists

[pabloarosado]

Written by Claude Fable 5 — @pabloarosado at the wheel.

Placeholder branch. Nothing implemented yet.

Problem

Three garden steps independently decide which country-years have no nuclear power, and two of them hardcode the answer:

• energy_institute/2026-06-30/statistical_review_of_world_energy: a 72-country list plus six hand-coded partial histories (Italy after 1990, Lithuania after 2009, Belarus from 2020, ...), around 150 lines with citations to external references.
• energy/2026-06-30/energy_mix: a separate 114-country list, covering the countries that only reach us through the EIA total extension, plus 5 known-nuclear exceptions.
• energy/2026-06-30/electricity_mix: no list, but it re-derives the zeros from the Statistical Review's already-filled data.

The two lists overlap on a single entity, so they are complementary rather than redundant. But both are maintained by hand and go stale whenever a country starts or stops generating nuclear power.

Proposal

EIA's international_energy already reports this as data. Its energy_consumption_from_nuclear column (terawatt-hours) covers 256 entities, including 131 of the Statistical Review's 132 (all but Curacao) and every historical state it reports. It agrees with every hand-curated entry:

• All 114 countries in the energy_mix list report exactly 0, and all 5 exceptions report nonzero.
• No contradiction with any of the 72 countries in the Statistical Review list.
• All six partial histories match, at the exact cutoff year.

Add a small garden step derived from EIA that reports nuclear status per country-year, then load it in the three steps above and delete their bespoke logic.

Constraints

• Three states, not two: "has nuclear", "no nuclear", and "unknown". An EIA data gap is not a zero, and collapsing the two would invent zeros.
• Use EIA's zeros only, never its nonzero values. A zero is methodology-independent, but a nonzero EIA figure sits on EIA's own accounting basis rather than the Statistical Review's total energy supply basis.
• This gives the Statistical Review a dependency on EIA for the first time, tying its build to another provider's release cycle.
• Curacao is absent from EIA and will still need an explicit answer.
• This touches the same function as 🐛 Energy mix: South America shares mix two producers, so it should land after it.

[owidbot]

Quick links (staging server):

Site Dev	Site Preview	Admin	Wizard	Docs	Docs Preview

Login: ssh owid@staging-site-refactor-nuclearstatus-eia

chart-diff: ✅

No charts for review.

data-diff: ✅ no differences

No differences found.

Automatically updated datasets matching excess_mortality|covid|fluid|flunet|country_profile|garden/ihme_gbd/2019/gbd_risk are not included. Run locally with etl diff REMOTE data/ --include <dataset> --verbose.

Edited: 2026-08-18 11:32:11 UTC
Execution time: 16.28 seconds