Curated Multi-Domain Energy and Socio-Economic Dataset for a 12-Country ENTSO-E Power System

This curated dataset integrates publicly available, multi-domain data representing the energy systems and socio-economic context of 12 ENTSO-E countries. The original data are neither generated nor owned by the author(s); rather, they are systematically aggregated from the following open-access sources: Electricity and Grid Data: Load, solar PV generation, wind generation, and cross-border power flows obtained from the ENTSO-E Transparency Platform. Meteorological Variables: Hourly atmospheric variables (e.g., air temperature, pressure) derived from ERA5 reanalysis data on pressure levels. Socio-Economic Indicators: Country-level population and GDP per capita sourced from the World Bank Data360 database. The dataset has been harmonized and structured to support research on spatio-temporal graph neural network (STGNN)–based load forecasting. Preprocessing procedures (e.g., temporal alignment, spatial aggregation, feature engineering, and normalization) are study-specific and tailored to the experimental design of each application. For methodological details and preprocessing specifications, refer to: Grid-Aware Probabilistic Residual Load Forecasting with Regime-Aware Conformal Calibration.