Elpriser/denmark-power-market

Denmark Power Market Data

Part of a country-split collection of historical electricity market data gathered for the elpriser.org price-forecasting project. See the index dataset for the other countries: Denmark, Germany, Norway, Sweden, Finland, Netherlands.

License & attribution

CC BY 4.0. EDS files (eds_*): Source: Energinet (www.energidataservice.dk). ENTSO-E files (entsoe_*): Source: ENTSO-E Transparency Platform (transparency.entsoe.eu).

Contents

EDS (Energinet, Danish TSO) — DK1 (West) and DK2 (East)

File	Description	Coverage
eds_dayaheadprices_dk1/dk2.parquet	Day-ahead spot prices (current dataset)	~2025-09-30 → present
eds_elspotprices_dk1/dk2.parquet	Day-ahead spot prices (legacy dataset, longer history)	2000-01-01 → 2025-09-30
eds_productionconsumptionsettlement_dk1/dk2.parquet	Actual generation by source and consumption	~2004-12-31 → present
eds_co2emis_dk1/dk2.parquet	CO₂ emission intensity of consumed electricity, g/kWh. Energinet publishes this at 5-minute resolution; averaged to hourly here so it joins directly onto the price files	2017-01-01 → present
eds_consumption_profile_private.parquet	Hourly electricity consumption of Danish private households (MWh). The "Privat" category of ConsumptionDK3619IndustryHour. National totals — not split by price area	2021-01-01 → present
eds_datahubpricelist.parquet	Nettariffer (every net company's grid tariff), Energinet's own system/transmission tariffs, elafgift (electricity tax), and subscription fees — the full DatahubPricelist price-component list, all charge owners, not split by DK1/DK2 (GLN_Number identifies the charge owner)	2014-01-01 → 2027-03-31 (includes announced future tariffs)

ENTSO-E — DK1/DK2 direct (complements the EDS files above)

File	Description
entsoe_dk1/dk2_load_actual.parquet	Actual (realised) total system load, MW
entsoe_dk1/dk2_load_forecast.parquet	Day-ahead forecast total system load, MW
entsoe_dk1/dk2_windsolar_forecast.parquet	Day-ahead wind+solar generation forecast (only ~1 day ahead, archived history)
entsoe_dk1/dk2_installed_capacity.parquet	Installed generation capacity per type, yearly snapshots

Interconnector data (Denmark's actual physical borders)

File pattern	Pairs covered
entsoe_flow_{pair}.parquet	Actual physical cross-border flow (MW), both directions
entsoe_ntc_{pair}.parquet	Day-ahead estimated Net Transfer Capacity (MW), both directions — sparse/empty for pure Nordic-Nordic borders (dk1_no2, dk1_se3, dk2_se4), which use flow-based (JAO) capacity allocation instead

Pairs: dk1_delu, dk1_no2, dk1_se3, dk1_nl, dk2_delu, dk2_se4.

Note on the CO₂ and household-consumption files

These two were added to support research on when electricity is cheap and clean, and whether consumers respond to it.

Two things they make visible in the Danish data:

• Carbon intensity has fallen sharply — the DK1 hourly average went from ~228 g/kWh in 2017 to ~85 g/kWh in 2025.
• Cheap is not the same as clean. The cheapest hours are increasingly the sunny midday hours, which can coincide with carbon-heavy imports. Joining eds_co2emis_* onto the price files hour by hour shows the two objectives diverging on many days.

The household profile is the standard weighting used for "profile prices" (a consumption-weighted average price, which reflects a real household bill better than a flat hourly mean). Because it is a national total, pair it with a national or area-weighted price series rather than a single price area.

Two data caveats, passed through unchanged from the source: the CO₂ series occasionally goes negative in DK1 (Energinet's methodology covers consumed electricity net of trade, and heavy wind exports can push the figure below zero), and roughly 55 hours across the nine years are missing, mostly DST transitions.

Note on eds_datahubpricelist.parquet

Columns include ChargeOwner, GLN_Number, ChargeType (D01 subscription, D02 fee, D03 tariff), Note/Description, ValidFrom/ValidTo, VATClass, Price1–Price24 (hourly for ResolutionDuration=PT1H tariffs, single monthly/yearly value otherwise), and ResolutionDuration. To reconstruct "nettarif C" for a specific net company, filter ChargeType == "D03" and Note == "Nettarif C" for that company's GLN_Number; Energinet's own system/transmission tariffs and elafgift are under GLN_Number = "5790000432752".

Note on Nordic flow-based congestion data

JAO's Nordic flow-based market coupling data (PTDF/shadow-price congestion data) is not included here or anywhere in this collection — JAO's terms of use prohibit redistribution without prior written authorization. If you want it, register for a free personal token at publicationtool.jao.eu.

Format

Parquet, one row per timestamp (and per production type / charge line, where applicable). Column names are preserved as returned by the source API, except in eds_co2emis_* where the 5-minute CO2Emission column is aggregated to hourly CO2Emission_g_per_kWh.

Provenance

Collected via each provider's public API using free, personally-registered access. Price, production, CO₂ and household-consumption files refreshed 2026-07-28; ENTSO-E load/capacity/interconnector files collected 2026-07-03/04. Collection scripts: elpriser.org repo, scripts/data_backup/ (fetch_all_history.py eds, eds-co2, eds-consumption, eds-tariffs).

A live JSON API over the same sources — current price, 7-day forecast, hourly CO₂ — is available at elpriser.org/api, no key required.