The rapid roll-out of smart meters in Scandinavia has enabled granular energy monitoring across building stocks, yet the tools to translate this data into actionable insights remain limited. To address this gap, this work presents Building-PROFet, a modular and scalable framework for estimating hourly heat and electricity demand profiles in buildings. The model builds upon Energy Signature (ES) techniques, introducing physically interpretable building parameters and extending them with periodicity adjustments and additional weather variables, such as wind speed and solar irradiance. A clear separation between estimation and prediction stages enhances model transparency and supports robust forecasts across temporal scales. Building-PROFet operates on real-world hourly smart-meter datasets, capturing diverse building types and energy efficiency levels. It supports aggregation to the neighbourhood or city level, allowing percentile-based scenario analysis for infrastructure planning and demand flexibility assessment. Validation against out-of-sample datasets from apartment blocks and commercial buildings demonstrates high predictive accuracy ( R 2 up to 0.98), particularly for heating demand. By capturing temporal dynamics and occupant-driven variability, Building-PROFet offers a powerful alternative to the traditional simulation-based approaches. It provides a reliable, interpretable energy demand profiles while ensuring privacy through parameter-based prediction. The tool is designed to assist engineers, energy providers, and policymakers in improving demand forecasting, optimising energy systems, and supporting risk-aware decision making. • Building-PROFet improves hourly heat and electricity demand forecasting methodology. • The model integrates weather, occupancy, and periodicity for accurate forecasts. • Validated on residential and non-residential buildings, achieving R 2 up to 0.98. • Supports energy system planning, peak load reduction, and grid flexibility strategies. • Provides insights into diversity of building stock’s energy demand.
Smertinas et al. (Wed,) studied this question.