This comprehensive CSV dataset provides metadata for 1,636 buildings, facilitating advanced energy analysis and performance benchmarking. The Building Data Genome Project 2 metadata set offers a structured repository of physical and operational characteristics for over 1,600 non-residential and residential structures. It includes essential fields such as site location, primary building use, gross floor area, year of construction, and industry classification. This data is intended to be joined with corresponding meter time-series data to enable deep-dive analysis into energy consumption patterns. By providing a standardized schema across a large, diverse sample size, it serves as a critical foundation for researchers and engineers developing predictive models or energy efficiency strategies.

The download is a single file containing .CSV files.

Professionals can integrate these CSV files into data analysis software such as Python, R, or specialized energy modeling platforms. By using the unique building identifier, users can map this metadata to primary time-series meter data, allowing for complex multi-variable analysis. The dataset is structured to facilitate seamless data cleaning and merging, enabling users to filter by specific building typologies, regions, or operational profiles to create targeted benchmarks for their own building performance simulation and optimization projects.

This dataset is primarily utilized by energy auditors, sustainability consultants, and building scientists working on large-scale portfolio optimization or academic research. It is invaluable during the early stages of benchmarking, where practitioners need to compare their subject buildings against a standardized, diverse sample. Engineers can use this information to inform energy baselines, identify outliers in utility consumption, or validate performance targets in commercial, institutional, and industrial projects. It pairs effectively with climate data sets and building energy simulation tools for comprehensive performance modeling.