The Building Data Genome Project Hourly Electricity Meters

The Building Data Genome Project provides a comprehensive CSV dataset of hourly electricity meter readings for 507 non-residential buildings. This dataset comprises one full year of hourly electrical meter data, structured as a benchmark collection for rigorous performance analysis. It encompasses a diverse range of 507 non-residential building types, providing granular time-series information essential for energy research. The files are provided in a standard CSV format, ensuring high compatibility with common data science and building performance modelling software. The collection serves as an empirical foundation for testing machine learning algorithms, validating energy simulation models, and establishing baseline consumption patterns across various institutional, commercial, and public facility typologies in a controlled, research-oriented format.

The download is a 53.3 MB file containing .CSV files.

Professionals integrate this dataset into analysis workflows by importing the CSV files into Python, R, or dedicated building energy management software. The data is typically used to train predictive algorithms, perform load profile clustering, or benchmark existing building performance against the provided empirical set. Users should ensure their environment is configured to handle large time-series data structures, typically mapping the hourly timestamps against specific building metadata to facilitate accurate correlation and comparative energy modelling during the design or operational stages.