Consolidated and standardized electrical panel and building characteristics in U.S. single-family buildings for predictive modeling
Data files
Mar 11, 2026 version files 4.81 MB
-
data_for_panel_modeling.csv
4.80 MB
-
README.md
3.71 KB
Abstract
This dataset includes building characteristics, panel capacity, and avaliable breaker space for single-family homes aggregated from field data collected by the TECH Clean California program, the Northwest Energy Efficiency Residential Building Stock Assessment, Home Energy Analytics, the Bay Area Renewable Energy Network, the Minnesota Center for Energy and Environment, the Berkeley Lab Citizen Science Survey, and the California Public Utilities Commission.
https://doi.org/10.5061/dryad.rv15dv4jw
Description of the data and file structure
This file contains data on building characteristics in single-family homes from multiple field data sources. The characteristics include: the year/decade of construction; primary fuels for space and water heating; cooking, and clothes drying; electrical panel capacity; available breaker spaces in the electrical panel; the presence of PV; the type and presence of cooling technologies; and the number of major electrical loads.
Files and variables
File: data_for_panel_modeling.csv
Description:
Variables
- panel_amp_pre: Report capacity (in Amps) for existing panels in single-family homes
- panel_amp_pre_bin: Binned panel capacity
- panel_slots_empty: The number of available breaker spaces (panel slots) in existing panels
- geometry_building_type_recs: Building type, aligned with definition used by the Residential Energy Consumption survey
- geometry_building_type_recs_simp: Simplified building type, collapses attached and detached homes
- year_built: Year of construction
- vintage: Decade of construction
- sqft: Floor area of building (in square feet)
- heating_fuel_simp: Primary space heating fuel. Non-electric includes gas, propane, fuel oil, and wood
- water_heater_fuel_simp: Primary water heating fuel. Non-electric includes gas, propane, fuel oil, and wood
- cooking_range_simp: Primary cooking fuel. Non-electric includes gas and propane
- clothes_dryer_simp: Primary clothes drying fuel. Non-electric includes gas and propane
- major_elec_load_count_w_ev_pv: Count of major electric loads (space and water heating, cooking, clothes drying, rooftop PV, cooling, and electric vehicle charging)
- has_pv: Presence of rooftop PV in home
- hvac_cooling_type: Type of cooling technology present
- has_cooling: Presence of cooling technology in home
Interpretation of 'null' values
For each variable, we have used 'null' as a placeholder for values not reported in the raw data that we used to construct this dataset. Users can interpret the 'null' values as unavailable.
Code/software
None.
If opened in a programming language, such as Python, be sure to permit 'None' as a non-null value. The Pandas Python package can deafult to reading 'None', which occurs in the drying and cooking fuels columns, as a null value.
Access information
Data was derived from the following sources:
- California Public Uitlities Commission. 2023 Viable Electric Alternative [dataset]. Published online March 5, 2024. https://www.cpuc.ca.gov/-/media/cpuc-website/divisions/energy-division/documents/building-decarb/fs-infra-ms-data-tool_draft_ver2.xlsx
- Northwest Energy Efficiency Alliance. 2016-2017 Residential Building Stock Assessment [dataset]. Published online 2017. Accessed November 14, 2022.
- Meier A. Berkeley Lab Citizen Science Survey [dataset]. Published online 2023.
- Home Energy Analytics. Home Energy Analytics [dataset]. Published online 2024.
- TECH Clean California. TECH Working Data Set [dataset]. Published online August 23, 2023. Accessed February 5, 2024. https://techcleanca.com/public-data/download-data/
- Jones K, Olson R, Otalora-Fadner A, Quinnell J. Minneapolis 1-4 Unit Residential Weatherization and Electrification Roadmap [dataset]. Published online 2023.
- BayRen. BayRen Home Energy Score and Electrification Checklist [dataset]. Published online 2022.
We standardized categorical field values in our dataset to account for variation in naming conventions across our datasets (e.g. ‘heat pump’ and ‘HP’). We also excluded outlier panel capacity values. We also screened out panel capacities that we suspected were data entry errors due their being unique to the dataset, non-integer, or not divisible by five. . We further excluded homes with panel capacities that were non-integer or not divisible by five as they did not agree with our understanding of electrical panel sizing. Additionally, we removed three homes less than 1,500 ft2 that had 10,000A panels, which we considered unrealistic. We also excluded homes with 100 or more reported available breaker spaces, which we did not find basis for in market-available electrical panels