Coverage

The four sources disagree about which countries exist, which years they cover, and how far forward they run. That is why every chart on the other pages filters for the columns it needs, and why those filters are not interchangeable.

How many countries each series actually covers

Renewables share stops at 79 countries where the electricity mix covers about 210, primary energy drops by two thirds in its latest year, and the series end in different years. Each has bitten a query here, and each is why a chart filters for exactly the columns it needs.

Two different kinds of gap

The fact is built on a spine, marts.dim_country_year: every country crossed with every year. Left-join the fact onto it and a gap comes back as a row you can count.

A country-year no source reports at all

Almost all are the deep past: the spine starts in 1750 because OWID's emissions series does. From 1990 on it is essentially complete, so this gap is about history, not the pipeline. These rows exist only in dim_country_year.

A country-year in the fact where one source is silent

These reach the fact with nulls where the silent source would be, and a where clause on the wrong column drops them without saying so.

World Bank data, no OWID emissions: small territories OWID does not estimate for.

OWID emissions, no World Bank GDP: absent from every intensity and decoupling measure on this site. Taiwan, not a World Bank member, emits more than the Netherlands.