hsilvosa/omie-electricity

OMIE: Iberian Wholesale Electricity Market (Spain and Portugal)

Standardized dataset containing over 17.15 million rows of market-clearing data from the Iberian wholesale electricity market, operated by OMIE (Operador del Mercado Ibérico de Energía).

It includes full aggregate supply and demand bidding curves block-by-block (curva_pbc) and the continuous time series of hourly marginal prices and cleared energy volumes.

---

Dataset Summary

Configuration	Rows	Temporal Coverage	Description	Parquet Size
bidding_curves	17,093,765 rows	Full Year 2024	Full supply (sale) and demand (purchase) matched and unmatched bidding steps per hour	210 MB
marginal_prices	54,983 rows	2023–2026 (Hourly)	Hourly marginal prices (EUR/MWh) and cleared energy (MWh) for Spain and Portugal	880 KB

---

Data Structure and Schema

1. bidding_curves (omie_bidding_curves_2024.parquet)

Discrete bid steps submitted by market participants (generators, retailers, and consumers):

Field	Type	Description
date	string	Market delivery date (YYYY-MM-DD).
hour	int32	Hour of the day (1 to 24/25).
curve_type	string	Type of curve (Venta for supply / Compra for demand).
unit_type	string	Unit classification (Nacional, Importación, Exportación).
energy_mwh	float64	Energy volume bid in this discrete step (MWh).
price_eur_mwh	float64	Bid price (EUR/MWh).
cleared_status	string	Clearing result for this bid step (Casada for cleared / No casada for unmatched).

2. marginal_prices (omie_hourly_marginal_price.parquet)

Continuous hourly time series resulting from the Day-Ahead market clearing algorithm:

Field	Type	Description
date	string	Delivery date (YYYY-MM-DD).
hour	int32	Hour of the day (1 to 24).
price_spain_eur_mwh	float64	Hourly marginal price for the Spanish bidding zone (EUR/MWh).
price_portugal_eur_mwh	float64	Hourly marginal price for the Portuguese bidding zone (EUR/MWh).
energy_spain_mwh	float64	Total cleared energy in Spain (MWh).
energy_portugal_mwh	float64	Total cleared energy in Portugal (MWh).

---

Usage

With Python (pandas / polars / duckdb):

import pandas as pd

# Load hourly marginal prices
df_prices = pd.read_parquet("data/omie_hourly_marginal_price.parquet")
print(f"Price range: {df_prices['price_spain_eur_mwh'].min()} to {df_prices['price_spain_eur_mwh'].max()} EUR/MWh")

# Load bidding curves (17M rows)
df_curves = pd.read_parquet("data/omie_bidding_curves_2024.parquet")
curve_hour_12 = df_curves[(df_curves["date"] == "2024-06-15") & (df_curves["hour"] == 12)]
print(f"Bidding steps in hour 12: {len(curve_hour_12)}")

---

Applications and Research Use Cases

• Reinforcement Learning (RL) and Battery Storage (BESS) Arbitrage: Simulating and training RL agents for optimal charge/discharge cycles and co-located storage dispatch.
• Short-Term Day-Ahead Price Forecasting: Machine learning models predicting hourly marginal prices and price spikes.
• Merit Order Dynamics and Renewable Impact: Empirical analysis of supply curve shifts due to solar PV and wind generation.
• Zero and Negative Price Modeling: Probability modeling of renewable curtailment and market decoupling between Spain and Portugal.

---

Provenance and Attribution

• Source: Operador del Mercado Ibérico de Energía (OMIE).
• Terms of Use: Publicly available data published by OMIE subject to source citation and attribution.