The gas-power nexus: why you can't model electricity in a vacuum

In European power markets the clearing price is set by the merit order, and at the margin — for most trading hours — sits natural gas. If you are forecasting day-ahead electricity prices or the frequency of winter spikes, grid fundamentals (load and wind) are only half the equation. You have to model the gas side too.

The alignment nightmare: zones vs countries

A proper gas-to-power feature set needs three ingredients:

1. Gas storage fullness and withdrawals — how much fuel is available. 2. Gas-fired electricity generation — how fast it is being burned. 3. Day-ahead power prices — what the market is clearing at.

The data exists — Gas Infrastructure Europe publishes storage (AGSI) and LNG (ALSI); ENTSO-E publishes generation and prices. The problem is spatial alignment. ENTSO-E data is by bidding zone: Germany and Luxembourg share DE_LU, Italy fractures into six zones, Denmark splits into DK1/DK2. Gas storage is published by ISO country code. So you cannot just JOIN German gas withdrawals to German gas-fired burn — you have to build a crosswalk that rolls bidding-zone power up to the ISO country, handling interconnectors and fractional mappings without double-counting.

The pre-joined solution

We built that mapping architecture once. Our pipeline reconciles GIE storage and LNG inventory with ENTSO-E generation and pricing at the ISO-country level, into clean daily panels:

• European Gas-Power Nexus — the pre-joined gas-to-power panel; load it and run your regressions (€19).
• European Gas Storage and LNG Inventory — daily AGSI / ALSI, cleaned.
• European Gas Balance Panel and Cross-Border Flows — supply and transmission.
• European Gas Flows — physical pipeline flows.

Don't spend a week mapping bidding zones to country codes. The panel is already joined — browse the gas & power catalog and start from the analysis.