Home › Methodology
Methodology

How we build the data.

Fraktix isn't scraped guesswork. It's an engineered pipeline over official government registries — the same machine we point at custom client datasets.

Sources: FMCSA · DVSA · SITR · refreshed monthly

The pipeline

1 · Official sources only

FMCSA (Company Census + Licensing & Insurance + Out-of-Service), DVSA Vehicle Operator Licensing (UK), SITR Registre des marchandises (France). Public filings, not scraped data.

2 · One normalized schema

Each registry defines "fleet" differently (power units vs licence copies vs authorised vehicles). We map them all to one clean, query-optimized record.

3 · Deduplication

One row per legal entity, keyed on official identifiers (DOT, SIREN, operator licence). Establishment- and director-level duplicates collapsed; corporate groups linked.

4 · Compliance & geo joins

Authority, insurance-on-file and out-of-service joined by DOT; every carrier geocoded to coordinates for proximity and territory.

5 · Lifecycle & freshness

A monthly full reconcile adds newly-registered carriers and retires removed ones — never silently deleted. Fleet-size and status changes tracked over time.

6 · Verification

Every record carries its source and an "as-of" date. Enrichment is confidence-scored and re-checked, not assumed.

Data quality & how we verify it

Official registries are self-reported and unvalidated — a small number of carriers file impossible figures, and others fall out of date. We don't simply republish them. Every record is cross-checked against correlated signals, implausible values are flagged, and rankings show only verified figures — with an "as-of" date on every field. We'd rather show a smaller, correct number than a bigger, wrong one.

Cross-field checks

Fleet size is validated against driver count and crash/inspection exposure. A carrier reporting 599,994 power units against 2 drivers is flagged, not published.

Independent references

Our largest-carrier lists are checked against independent industry rankings, so a "top carriers" list reflects reality — not a data-entry typo that outranks the real leaders.

Source & freshness on everything

Every field carries its official source and a last-updated date, re-checked on the monthly refresh. Confidence is scored, not assumed.

Frequently asked questions

Where does Fraktix's data come from?

Only official government registries: the FMCSA Company Census and Licensing & Insurance system (US), DVSA Vehicle Operator Licensing (UK), and the SITR Registre des marchandises (France). No scraped or purchased mystery data — every field is traceable to a public filing.

How often is the data refreshed?

Monthly. A full reconcile adds newly-registered carriers and retires removed ones — records are never silently deleted, and status and fleet changes are tracked over time. Every record carries the source and an 'as-of' date.

How many carriers does Fraktix cover?

About 2.33 million active carriers across five countries — 2.2M in the US (FMCSA), 64K in the UK (DVSA) and 48K in France (SITR), plus cross-border Canadian and Mexican carriers operating under a US DOT number.

How are duplicate records handled?

One row per legal entity, keyed on official identifiers — DOT number (US), SIREN (France) or operator licence number (UK). Establishment- and director-level duplicates are collapsed and corporate groups linked.

What compliance data is included?

For US carriers, the free FMCSA compliance cluster: operating authority (common/contract/broker), insurance-on-file (BIPD) with minimum coverage, and out-of-service status — joined to each carrier by DOT number.

Can I get contact details or custom enrichment?

The public directory shows registry facts. Verified role-based contacts, insurer, safety and renewal signals, and any custom fields are delivered through Data Services — built to your ICP, verified, and refreshed on your cadence.

How do you handle errors in the government data?

The registries are self-reported and unvalidated, so a small number of records are inflated (we've seen a carrier file 599,994 power units against 2 drivers) or out of date. We run cross-field consistency checks — fleet size against driver count, and crash and inspection exposure — to flag implausible values, validate our largest-carrier lists against independent references, and surface only verified figures in rankings. Every record carries its source and last-updated date. We would rather show a smaller, correct number than a bigger, wrong one.

Why might a carrier's fleet size look low or unexpected?

Fleet size comes from the carrier's own MCS-150 filing, which they update periodically — so a fast-growing carrier can lag reality until its next update, and occasional typos slip into the source. Our monthly refresh keeps figures current, and we flag records whose fleet size disagrees with their driver count or operating history.

Data Services
This pipeline can be pointed at your use case.Custom fields, custom targeting, verified enrichment and delivery — API, CSV, or a database you own.
Build a custom dataset →