This dataset contains simulated hourly end use load profiles of the residential and commercial building sector in the contiguous United States for every other year from 2010 to 2050. Data were produced in 2021 using ResStock and ComStock, which are building stock energy models of the US residential and commercial sector, respectively, and are published in dsgrid Toolkit format.

The dataset consists of base year 2018 ResStock and ComStock (collectively known as BuildStock) timeseries data differentiated by county, building type, fuel type, and end use, along with backward-and forward-looking projections created by applying regional-, sectoral-, and end use-specific growth rates derived from EIA's 2021 Annual Energy Outlook (AEO)'s Reference Scenario. The base year datasets represent the US building stock as of 2018 and were simulated in 2021 using AMY 2012 weather to align with NREL's wind and solar resource datasets. They were produced using the BuildStock tools during the End Use Load Profiles (EULP) calibration project. The projection methodology is described in the technical report linked below. Reflecting EIA's reference scenario assumptions to provide a baseline for exploring long-term trends, the projection does not reflect large-scale electrification of building space heating, water heating, clothes drying, cooking, or other end uses. The dataset also does not include electric vehicle charging that might occur on-site at buildings. Electric vehicle charging is described in the dsgrid TEMPO Light-Duty Vehicle Charging Profiles v2022 (see "dsgrid TEMPO" link below).

This dataset describes a reference projection of building energy consumption at a resolution sufficient for bulk power system and other forms of regional energy system planning. It improves on traditional load forecasting practices in the power sector by providing annual hourly data resolved geographically, temporally, and sectorally using state-of-the-art sector-specific energy modeling tools and dimensionally aligned (i.e., regionally, sectorally, and end-use specific) growth rates. Compared to previous practice of regional load forecasts using a single load shape and all-electricity growth rates, the product is a more resolved dataset that is easier to align with the geographic resolution of power sector production cost and capacity expansion models and more capable of representing load shape changes induced by uneven growth across sectors or technology types. The parameterization of the growth rates could also enable creation of alternative scenarios with different amounts of electrification and energy efficiency.

The full dataset as well as various aggregations are available for access. Large datasets are in parquet format, with some partitioned by a few key dimensions. Smaller datasets are available as csv.