When we first built the EPA-to-FMCSA carrier matching system inside Enviro Logistics, we used three signals: company name similarity, shared phone number, and geographic proximity. That got us to 1,346 FMCSA hazmat carriers we could confidently call EPA-registered hazardous waste transporters.
We knew the number was incomplete. The EPA RCRAInfo database has 12,303 active transporter records. We had matched 4,799 of them, or 39%. The other 7,504 EPA transporters had no FMCSA counterpart we could find. Three months passed. The problem sat there.
This week we went back in.
What Was Blocking the Other 7,504 Matches?
Name matching has a punctuation problem. “NRC ENVIRONMENTAL SERVICES INC” and “NRC ENVIRONMENTAL SERVICES, INC.” are the same company. But a trigram similarity algorithm treats the comma and period as characters, which reduces the similarity score just enough to fall below the matching threshold. The same issue applies to ampersands: “NATURE ENVIRONMENTAL & MARINE SERVICES” versus “NATURE ENVIRONMENTAL AND MARINE SERVICES” is an identical company with a different registration convention, and the original algorithm returns zero match.
Phone matching has a state problem. If a carrier’s DOT registration lists a corporate phone number in Texas, and their EPA registration lists the same number for a facility in Louisiana, the original phone matcher rejected it as a cross-state false positive. Those rejections were correct at the time. They also meant legitimate multi-state carriers got missed.
Name fuzzy matching has a volume problem. When you run trigram similarity at a low threshold across 150,000 FMCSA carriers and 12,000 EPA handlers, you generate millions of candidate pairs. Most are noise. “A A A TRANSPORT INC” matching “AERO PATH TRANSPORT INC” because both contain the word “TRANSPORT” and end in “INC” is not a useful match. We had already cleaned 60,000+ false positives from the first pass.
The Three Signals We Added
Signal 1: Shared corporate email domain. We filtered out generic providers (Gmail, Yahoo, Hotmail, Comcast, and 14 others) and matched on the remaining business email domains within the same state. A trucking company and an EPA handler sharing @acmetruckline.com in the same state is a strong signal. We required name similarity of at least 40% as a second confirmation. This produced 162 new matches, 114 of which were active FMCSA hazmat carriers that had never shown an EPA badge.
Signal 2: DBA name matching. FMCSA records 35,656 hazmat carriers with a “doing business as” name that differs from their legal name. The EPA may have registered the company under that DBA. We normalized both names (removed punctuation, expanded ampersands) and ran an exact match between EPA handler names and FMCSA DBA names within the same state. This produced 26 new matches, clean and high-confidence, such as “CTSI LOGISTICS” (DBA) matching the EPA record for the same company registered under its full legal name.
Signal 3: Physical address proximity. Both the EPA and FMCSA record GPS coordinates for registrants. We matched carriers located within 5 kilometers of an EPA transporter site and required name similarity of at least 70% as confirmation. This produced 71 new matches with no ambiguity. The top results were companies like “GABEL’S HAULING AND DEMOLITION INC” exactly matching “GABEL’S HAULING AND DEMOLITION INC.” at the same GPS coordinates, where the original algorithm failed purely because of an apostrophe and a period.
The Result
1,346 EPA-registered hazmat carriers on the platform before. 1,556 after. That is 210 freight operators who are verified EPA hazardous waste transporters, now visible to brokers and shippers who were looking for them.
The most interesting thing about this upgrade is what it reveals about the underlying data problem. The FMCSA and EPA are separate federal agencies with separate registration systems. There is no official linkage between a DOT number and an EPA Site ID. Every carrier that operates in both systems registered independently, often years apart, sometimes under slightly different legal names or with different contact details on file. The matching problem is not a database problem. It is a records-reconciliation problem, and the most powerful signals are not the obvious ones.
Company name is the obvious signal. Email domain, physical location, and DBA registration are the ones that fill the gaps.
We expect to continue expanding this as our EPA scraper completes national coverage (currently at 71.7%) and as we run additional normalization passes over the remaining unmatched EPA transporter records.
Search EPA-registered hazmat carriers at app.envirologistics.online.
Helpful Resources
- RCRAInfo Search (Envirofacts) — EPA’s public database of registered hazardous waste transporters referenced throughout this article.
- EPA Form 8700-12 — the Site Identification Form that puts a carrier into the RCRAInfo transporter registry in the first place.
- FMCSA SAFER Company Snapshot — the DOT-side hazmat authority record this matching work cross-references against EPA data.
