Open Datasets

Raw Building Metadata

Semantic metadata standards for buildings such as Brick and Project Haystack show promise in enablilng wide-scale deployment of energy-efficiency measures and advanced building management technologies. Converting existing building metadata to these standard forms is an area of active research.

The data available here is a collection of metadata pulled from real building management systems around the world. By making this data available, we hope to facilitate research into the automated conversion of unstructured metadata into standard forms. The dataset currently contains attributes for 103,064 points from 92 buildings.

Data is distributed in CSV form:

This dataset is not static!. Email CSV dumps of building metadata (or questions) to data [AT] mortardata [DOT] org. We have also produced an open-source tool for scraping and cleaning building metadata from existing buildings.

The dataset and data collection tool are described in our paper in the DATA 2019 workshop.

Mortar Testbed

Access to large amounts of real-world data has long been a barrier to the development and evaluation of analytics applications for the built environment. Open data sets exist, but they are limited in their span (how much data is available) and context (what kind of data is available and how it is described).

The goal of Mortar is to provide a large, diverse and consistently updated testbed of buildings and building data to facilitate reproducible evaluation of building analytics.

At this time, Mortar contains 107 buildings, spanning over 10 billion data points and 26,000 data streams. Context for these data streams is provided by a Brick model. The Brick model describes for each building: (1) what data streams exist and what they measure, (2) what equipment exists and how it is monitored, (3) the relationships between equipment in terms of flows, composition and location.

Information how to access the Mortar dataset is available at