BuildingCensus

Methodology & data

Every figure on BuildingCensus traces to a government's official building-energy-benchmarking disclosure. Nothing is modeled, estimated or scraped from third parties.

The unit of coverage

What this is a census of

The US has about 5.9 million commercial buildings (EIA CBECS 2018). Their energy use is private — except where a benchmarking-disclosure law makes it a public record. Roughly 50 US cities, counties and states have such laws, and they apply only to large buildings (thresholds of 20,000–50,000 sq ft are typical). BuildingCensus is the census of that disclosed record: every building published by every jurisdiction with public data, and an explicit coverage roadmap for the 22 jurisdictions whose data is queued or not yet public.

Sources

The 24 live disclosures

CityLaw / programBuildingsPortal
New York City, NYLocal Law 84 (LL84)48,310source →
Chicago, ILChicago Energy Benchmarking Ordinance3,852source →
Seattle, WASeattle Energy Benchmarking Ordinance (SMC 22.920)3,823source →
Los Angeles, CALA Municipal Code 22.610.1 (EBEWE)11,498source →
Cambridge, MABEUDO902source →
San Francisco, CASF Environment Code Chapter 202,498source →
Boston, MABERDO 2.06,305source →
Washington, DC, DCClean and Affordable Energy Act / BEPS6,870source →
Denver, COEnergize Denver Building Performance Policy3,287source →
Philadelphia, PAPhila. Code §9-3402 / Building Energy Performance Program2,379source →
Montgomery County, MDBenchmarking Law (Bill 2-14) / BEPS (Bill 16-21)1,766source →
Kansas City, MOEnergy Empowerment Ordinance (No. 150299)1,628source →
San José, CASan José Municipal Code Ch. 17.87 (BPO)1,790source →
San Diego, CABuilding Energy Benchmarking Ordinance (SDMC Ch. 14 Art. 12)2,567source →
Berkeley, CABESO (BMC Chapter 19.81)430source →
California, CAAB 802 (Public Resources Code §25402.10)40,029source →
Atlanta, GACommercial Buildings Energy Efficiency Ordinance (CBEEO)660source →
Ann Arbor, MIBenchmarking and Disclosure Ordinance (2021)418source →
Evanston, ILBenchmarking Ordinance (67-O-16)366source →
St. Louis, MOBuilding Energy Awareness Ordinance 70474 / BEPS 71132639source →
Miami, FLBE305 Benchmarking Ordinance434source →
Portland, ORCommercial Building Energy Performance Reporting Policy952source →
Honolulu, HIOrdinance 22-17 (City & County of Honolulu)799source →
Reno, NVReno Energy and Water Efficiency Program Ordinance233source →

In the queue (public data, bulk ingest pending): Massachusetts (statewide), Minnesota (statewide), Boulder, CO, Columbus, OH, Chula Vista, CA, Des Moines, IA, Portland, ME, Salt Lake City, UT, Hennepin County cities, MN (Edina, St. Louis Park…). Jurisdictions whose laws haven't produced public data yet are listed on the coverage page.

Identity

How a building is keyed

Cities publish one row per building per reporting year 538,417building-year records across the corpus. BuildingCensus keys each building by its city's own building identifier (BBL-derived IDs in New York, OSE building IDs in Seattle, parcel numbers in San Francisco, and so on). The latest reporting year is canonical— it drives the page's grade, stats and metadata — and every earlier year renders in the year-over-year table.

Profile URLs end in a stable 8-character fingerprint of that city-plus-building key, so a page's address never changes when a new disclosure year lands.

Honest data

Per-source caveats

  • New York City: The source lists the borough, not the city, in its own city field — BuildingCensus normalizes every record to New York City.
  • Chicago: Chicago reports electricity and gas only as kBtu totals; kWh and therm figures are unit conversions, not source-reported values.
  • Seattle: Seattle's dataset reports no water-use field.
  • Los Angeles: The LA source publishes no building names and no coordinates — pages are addressed by street address, and LA buildings do not appear on maps.
  • Los Angeles: LA reports EUI and emissions only (no electricity/gas breakout).
  • Cambridge: The Cambridge source publishes no building names — pages are addressed by street address.
  • Cambridge: No per-square-foot GHG intensity in the source (total emissions only).
  • San Francisco: SF classifies buildings into four broad categories rather than EPA property types.
  • San Francisco: Gas figures are kBtu-converted; no water-use field.
  • Boston: Boston publishes no building names or coordinates — pages are addressed by street address.
  • Boston: Emissions arrive as kgCO₂e and are converted to metric tons; 2021-and-earlier files use the old BERDO 1.0 schema and are not yet loaded.
  • Washington, DC: The DC source stores every metric as text — values are cast to numbers on ingest.
  • Washington, DC: Rows where a building didn't report that year carry no metrics (the disclosure itself is the record).
  • Denver: Denver publishes no coordinates in any final dataset — buildings don't appear on maps.
  • Denver: The in-progress reporting cycle is deliberately excluded; only final per-year master datasets (2021–2023) are loaded.
  • Philadelphia: Electricity and gas arrive only as kBtu totals — kWh and therm figures are unit conversions.
  • Philadelphia: 2013–2017 disclosures exist on a legacy portal with a different schema and are not yet loaded.
  • Montgomery County: Only the 2023 and 2024 disclosures carry energy metrics — earlier county files are compliance-status-only.
  • Montgomery County: Names and addresses arrive lowercase and are title-cased for display.
  • Kansas City: EUI figures are weather-normalized (the source publishes no raw EUI).
  • Kansas City: Total emissions are the sum of the source's direct and indirect GHG columns.
  • San José: The source publishes no building names or coordinates — pages are addressed by street address.
  • San Diego: Only the complete 2021 and 2022 cycles are loaded — the city marks 2023–2025 cycles as partial or under data-quality review.
  • San Diego: The public payload carries no source EUI, water or fuel breakout.
  • Berkeley: Berkeley publishes one wide table with per-year columns — years a building didn't report are omitted rather than shown as empty rows.
  • California: This is the statewide AB 802 disclosure — it overlaps the city programs of Los Angeles, San Francisco, San José, San Diego and Berkeley, which remain separate jurisdictions here because each is its own public record.
  • California: EUI figures are weather-normalized (the CEC publishes no raw site EUI).
  • Atlanta: Atlanta's public layer is a geocoded analysis dataset: site EUI and ENERGY STAR score only — no emissions, water or fuel figures.
  • Atlanta: Roughly 1,800 source rows carry no report year and are excluded.
  • Ann Arbor: Zeros in the source mean 'not reported' and are treated as missing values.
  • Ann Arbor: Coordinates are parcel centroids, not rooftop points.
  • Evanston: The city has published nothing after the 2019 data year — the newest disclosure on record is 2019.
  • Evanston: The 2019 file reports emissions in kgCO₂e (converted to metric tons on ingest).
  • St. Louis: The only public per-building file is the Cycle 1 report: weather-normalized site EUI for 2017–2021 — no ENERGY STAR scores, emissions or fuel figures.
  • Miami: Miami publishes no building names or coordinates — pages are addressed by street address.
  • Miami: Disclosures begin with reporting year 2023.
  • Portland: 2020–2022 reporting years were disrupted (COVID-era waivers) — the loaded record jumps from 2019-era files to 2023.
  • Honolulu: Covers all of Oʻahu (City & County of Honolulu), not just the urban core.
  • Reno: The city publishes a redacted file — some fields are blanked, and EUI figures are weather-normalized.
  • Columns a source doesn't report are left empty — we never fabricate a value to fill a table.

Quality

The completeness gate

Buildings that report neither an ENERGY STAR score nor a site energy-use intensity, or that are under 10,000 sq ft, still get a page (the record is public either way) but are marked noindex and excluded from sitemaps. Search engines and answer engines only see buildings with real performance data.

ENERGY STAR letter bands (A 85–100 / B 70–84 / C 55–69 / D 1–54) follow New York City's LL33 grade cutoffs and are applied uniformly across cities for comparability. They are BuildingCensus presentation, not an official rating outside NYC.

Use

Licensing & attribution

The underlying disclosures are public records published by each city. BuildingCensus's unification, grading, cross-links and history views may be cited with attribution to BuildingCensus (buildingcensus.com). For bulk or API access, contact kenny@hyder.me.

Dataset updated . Freshness reflects the most recent source re-ingest.