Methodology & data
TheFiscalFinder turns federal spending into a browseable graph. The graph is only worth anything if the links are trusted, so we label every fact by how we know it.
Where the data comes from
All figures come from USASpending.gov, the U.S. government’s official source for comprehensive federal spending data: awards, recipients, geography, and agencies. It is a work of the U.S. government and is in the public domain (17 U.S.C. §105). We currently load fiscal years 2023 to 2025.
How we label confidence
Every displayed fact carries one of three labels:
- ✓ from sourceFrom source. Taken straight from a federal filing: an award amount, an agency name, a recipient’s UEI, whether a contract was competed.
- ≈ computedComputed. Calculated by us from many source rows: totals, rankings, the “top recipients” lists, competitor groupings by industry code.
- ≈ inferredInferred. Derived with real uncertainty: a congressional district guessed from an address, a corporate parent the recipient self-reported. We word these carefully and never state them as fact.
What we cover in Phase 1
We profile the top recipients of federal awards by fiscal-year obligation (currently about 500 combined organizations) and the largest individual awards. Award pages whose recipient falls outside that set say so instead of linking to a page we don’t have. The award itself is still real and from the source. Coverage grows as the project does.
Congressional districts are inferred
The district where work is “performed” is derived from an address and is frequently imprecise or missing in the source. We always label district figures as inferred and validate them against a FIPS-to-district crosswalk. Treat them as estimates, not official tallies.
The political layer: what it is
Phase 3 adds a layer of public political data on top of the spending. The roster of Congress comes from the unitedstates/congress-legislators project, a public-domain dataset of current and historical legislators, committees, committee membership, and FEC IDs (the ids we use for matching). Campaign money comes from the Federal Election Commission, the official source for candidates, committees, and their raising, spending, filings, and contributions. Both are public record, and we add no private data.
A note on labels: where we simply repeat a filing, we mark it “from source.” Where we join, normalize, or curate that data (combining registrations, inferring a district, ranking totals), we mark the result “computed” or “inferred.” We do not present a transformation of the data as if the government had published it that way.
Who represents a place is on the record; the dollars are inferred
Which member represents a district or state is a fact from the congressional record, and we label it as such. What is inferred is which district a given dollar belongs to, because that is derived from a place-of-performance address that is frequently imprecise or missing. So “represented by” is reliable, but “federal work in this district” inherits the same inference caveat as every district figure on the site. We never treat spending in a place as something the member did.
Committee oversight is jurisdiction, not control
We map some agencies to the congressional committees whose public jurisdiction covers them, for example the health committees and the Department of Health and Human Services. These links are a curated reading of public committee assignments, so we label them inferred. Oversight is the authority to hold hearings and write law. It is not the power to direct a specific award, and we do not present it that way.
Political money: disclosed, regulated, and not a verdict
We show contributions that political action committees report to the Federal Election Commission, and link them, through the entities behind them, to the members who received them. The matching uses the FEC identifiers already carried in the congressional roster. Everything here is a public filing. Federal campaign contributions are regulated, subject to limits, and publicly reported. We show them as context, placed beside federal spending, so the public can see the whole picture. We never assert that a contribution caused a contract, a vote, or any official act, and a contribution appearing in FEC data is not a determination by us that any filing detail is lawful. Where we rank or total these figures, the ranking is a computation over public facts, not a finding of influence.
Corporate families: combined, and labeled
Large contractors register with the government under many identifiers (UEIs). When several registrations share the same registered name, we combine them into one page and label the result “combined across N registrations.” Each individual registration keeps its own drill-down page. Links that come from a recipient’s own filings (“ultimate parent”) are shown as “reported,” and we flag uncertain matches rather than over-merging companies that merely share a similar name.
Some well-known parents own subsidiaries under different names: General Dynamics owns Electric Boat, UnitedHealth Group owns Optum, and Lockheed Martin acquired Sikorsky in 2015. For a curated set of major parents we group those subsidiaries under the parent brand and combine their totals, drawn from public ownership records (SEC filings and official announcements). These groupings are labeled “≈ inferred · corporate family,” carry their source on the parent’s page, and never fold in a joint venture, which is a shared entity rather than a subsidiary. Every subsidiary keeps its own page.
How we resolve entities
A single company can appear under many name spellings and across the legacy DUNS-to-UEI transition. We merge conservatively, preferring exact identifier matches and flagging fuzzy ones, because a wrong merge is worse than a missed one.
Context, not accusation
Inferred links (a district, a corporate parent, a computed competitor, a PAC connection) are analytical conveniences, not assertions of influence or wrongdoing. Federal contracting and campaign contributions are normal, lawful, disclosed activity. TheFiscalFinder places public records side by side so you can follow the money; it does not follow the corruption.
No allegation of illegality, corruption, improper influence, or wrongdoing is intended regarding any named person, committee, agency, contractor, PAC, or entity. Every figure is labeled by how we know it, and inferred links are never stated as fact.
Corrections
Found something wrong? We’d like to fix it. Reach us via the about page. Corrections flow through a reviewed update, not a silent edit.