Epa Cdr Chemical Volume Data: Dataset Structure and Field Coverage

The EPA Chemical Data Reporting (CDR) Database holds what US manufacturers and importers must disclose under TSCA Section 8(a): production volumes in pounds for every reportable chemical, broken out at three levels - standardized parent company (domestic or foreign, with Dun & Bradstreet numbers), individual manufacture or import site (name, address, latitude, longitude, NAICS code), and chemical identity (CAS Registry Number or accession number).

For field-by-field documentation of all 27 columns, see the EPA Chemical Data Reporting (CDR) Database dataset page.

Four numeric columns split every headline number, so you can sum them or keep them separate:

(Column labels are cycle-specific; the 2020 cycle uses 2019 as its principal year.) Around the volumes sit 23 more fields across the 27-column schema: an ACTIVITY flag distinguishing domestically manufactured from imported chemicals, IND SECT CODE among 48 industrial sector options, INDUSTRIAL FUNCTION CATEGORY among 117 options with IND PV PCT percentages, consumer and commercial product categories spanning 97 options, physical forms drawn from seven options, worker-exposure counts reported as coded ranges, and USED IN PROD FOR CHILDREN.

One structural caveat shapes every analysis: processing and use information covers activities in the final year of each cycle only, not the whole reporting window. Confidential business information appears only as ranges or accession numbers, with NKRA marking fields that are not known or reasonably ascertainable.

Treat the four PV columns as the analysis surface and the rest as segmentation. Nationally aggregated PV is where EPA puts a usable per-chemical total when site-level detail is withheld, so national trend work survives confidentiality treatment even where plant-level ranking does not. The ACTIVITY flag then tells you which side of the domestic/import split each record contributes to - useful when import reliance rather than domestic manufacture is your question.

Three workflows recur, all starting from site coordinates:

• Market sizing joins CDR site rows to the Census Data API's 2022 Economic Census Basic feed (130 documented variables, county-level establishments, employment, payroll and revenue under commercial delivery terms), filtering NAICS to basic chemical manufacturing codes and weighting counties by CDR volumes.
• Supply-chain mapping rolls site-level pounds up through STANDARDIZED PARENT COMPANY NAME plus domestic and foreign D-U-N-S numbers to corporate footprints - the only public route to that view for US chemical manufacture.
• Exposure and ESG screening pairs CDR volumes with EPA Toxics Release Inventory (TRI) records at matched sites. Both report pounds, but they measure different things: CDR is how much a company makes or imports every four years; TRI is what facilities release or transfer annually, back to reporting year 1987, with fugitive and stack air splits and underground injection divided into Class I and Class II-V wells. Join on EPA Facility Registry ID and NAICS on the CDR side versus TRIFD and FRS ID on the TRI side.

Volume context also comes from the EPA TSCA Chemical Substance Inventory's July 2026 snapshot of 86,741 substances - 42,698 active against 44,043 inactive - which supplies the commercial-status screen for whether a reported chemical remains in commerce at all.

Within Datadory's catalog nothing else gives you named company-site-chemical volume records for the United States. The closest structural analogue is European and measures emissions rather than production: the EEA European Industrial Emissions Portal (E-PRTR) covers over 60,000 industrial sites with facility-level pollutant releases for 2007-2023, free under CC-BY 4.0 in 16 CSV extracts. Weigh the two in vs EEA European Industrial Emissions Portal (E-PRTR).

Against its EPA siblings: TRI scores the same 9 but captures releases, the TSCA Inventory lists substances without volumes, and ChemView aggregates per-chemical documents through a web app only. If your question needs annual granularity instead of four-year cycles, TRI is the companion; if it needs revenue-side market sizing down to counties, pair CDR with Census ecnbasic. All 27 pooled datasets for the sector sit in the commodity chemicals data guide.