Data Sources and Methodology

This page lists every dataset this site reads from, what is included and excluded and why, and the exact calculation behind every number shown on a contractor or hub page.

Data sources

DatasetPublisherLicenceUpdate frequency
USAspending Award Data Archive — ContractsU.S. Department of the Treasury (usaspending.gov)U.S. federal government work — public domain (17 U.S.C. § 105)Monthly (archive republished mid-month)
North American Industry Classification System (NAICS)U.S. Census BureauU.S. federal government work — public domainAs revised by the publisher
Product and Service Code (PSC) ManualU.S. General Services Administration (acquisition.gov)U.S. federal government work — public domainAs revised by the publisher

Contract data on this site was last downloaded 2026-09-18.

Scope

This site covers federal contracts only — award types A, B, C, D, and IDVs — for fiscal years 2021–2026. Grants, loans, and direct payments are out of scope, because those recipient categories include far more individuals and sole proprietors. The current fiscal year is always partial until it closes on September 30.

What is read from the source, and what is not

The pipeline reads a fixed, whitelisted set of columns from each USAspending contract transaction file — award identifiers, dates, dollar amounts, the recipient's legal name, Unique Entity ID (UEI), CAGE code, city and state code, parent-company name and UEI, business size, NAICS and PSC codes, contract-pricing type, and set-aside type.

It never reads a street address, phone number, e-mail address, or any free-text award/transaction description field, even though those columns exist in the source. They are excluded at the column level, before any privacy filtering runs — not filtered out after the fact.

Who is excluded, and why

A recipient is excluded entirely — no profile, no mention in any hub table — if any of the following is true for any contract action on file:

  • The award record itself flags the recipient as a sole proprietorship.
  • The recipient name is "Multiple Recipients," "Redacted Due To PII," "Miscellaneous Foreign Awardees," or otherwise contains "redacted."
  • The name contains no recognizable business or legal-entity word (LLC, Inc., Corp., and similar) and the recipient is classified as a small business — treated as a likely individual name rather than a company, since a false negative here would publish a person's name.
  • The recipient's UEI is on this site's manually maintained removal list.

On the current data (197,832 distinct UEIs across 6 fiscal years), this excluded 29,809 sole proprietorships and 11,268 likely individual names, leaving 156,753 recipients eligible for consideration.

Addresses are never shown more precisely than city and state — a street address is never read from the source at all, so there is nothing to redact at this step.

Publication threshold and scoring

A contractor qualifies for a profile page if, over the last 3 closed fiscal years, it has:

  • at least 3 distinct contract awards, and
  • at least $250,000 in combined net obligations, and
  • at least one contract action within the last 24 months.

Qualifying contractors are ranked by a score out of 1.00: 0.60 × percentile(3-year obligations) + 0.25 × percentile(3-year award count) + 0.15 × recency, where percentiles are computed across all qualifying contractors and recency is 1.0 if the contractor had activity in the last 12 months, else 0.5. The top 15,000 scores form the publication "wave"; a page is only built once a contractor is inside the wave.

On the current data, 37,748 contractors met the threshold; 15,000 are in the current wave (scores 0.6342–0.9998), and 3,000 of those have a published page so far — the wave is published in batches of 3,000, expanding over time rather than all at once.

If a parent company and one of its own subsidiary UEIs both qualify under the same normalized name, only the higher-scoring one gets its own page; the other is listed on that page as a related UEI under the same parent.

What each figure means

  • Obligations are net federal obligations — new obligations minus de-obligations — for the fiscal year(s) shown, not total contract ceiling value.
  • 3-year obligations / award count cover the last 3 closed fiscal years, not the current partial one.
  • Share (%) on an agency/NAICS/PSC row is that row's obligations divided by the contractor's total obligations across every fiscal year in the export.
  • "Similar contractors" are other published contractors with the same primary NAICS code, ordered by how close their total obligations are.
  • Hub pages (by agency, NAICS, state, and so on) list every published contractor in that group, largest first, capped at 500 rows.

Questions about a specific number, or a correction request, go through the contact form.