This dataset includes three years of cleaned hourly data from 3021 commercial smart heat meters installed in Danish residential buildings. The data are screened, interpolated to be equidistant, and missing values were imputed using a weighted moving average combined with a scaling algorithm to obey the data's cumulative trend. The original (anonymised) raw data of 3127 smart heat meters are also provided to increase transparency and reproducibility. Together with the consumption data, contextual information about the construction year, the type of building, and, if available, the energy label for the smart heat meters buildings (for all 3127 buildings of the raw data) are provided. The unique meter ID can link this data to the consumption data. A .pdf document describing the purpose of every data column is given in the folder '01_Data'. Besides this, three figures visualising the z-normalised data in different temporal resolutions are provided. Next to the data and the data visualisation, all code, written in R, used for data processing and extensive technical validation of the data is included. For a more extensive description, we would like to refer to our peer-reviewed open-access article describing the dataset and its processing: https://doi.org/10.1038/s41597-022-01502-3. Please also consider citing this article if you use this dataset. (Schaffer, M., Tvedebrink, T. & Marszal-Pomianowska, A. Three years of hourly data from 3021 smart heat meters installed in Danish residential buildings. Sci Data 9, 420 (2022). https://doi.org/10.1038/s41597-022-01502-3)

This work was funded by the independent Danish research fund number 0217-00340B. This project has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement No. 958345 (PRELUDE).