Data sources
Everything on this site is built from public records. These are the datasets we use, and what each one covers.
Candidate financial totals
How much each campaign raised, and how much of that came from individual donors, political action committees, and party committees.
openFEC: /candidate/{id}/totals/
Figures are reported by the campaigns themselves and filed with the Commission. We do not adjust them.
Independent expenditures
Money spent for or against a candidate by outside groups such as super PACs. It never passes through the campaign.
openFEC: /schedules/schedule_e/by_candidate/
We use the Commission's aggregated endpoint rather than summing individual rows. Raw rows include 24-hour notices and amended filings that would otherwise be counted twice.
Bundled (earmarked) contributions
Individual donations that an organization collected and directed to a named candidate. Legally individual money, so it never appears in a PAC total.
openFEC: /schedules/schedule_a/, the Commission's earmark attributions
Counted only for the organizations we have swept so far, which the methodology page lists. Repeated amounts from amended filings are removed before publishing.
Committee classifications
Whether each giving committee belongs to a corporation, trade association, labor union, or membership organization, and the parent organization's name.
FEC bulk data: committee master file (cm)
The classification is the Commission's own. The industry names layered on top of it are ours, and the full mapping is published on the methodology page.
Roll-call votes
Every recorded floor vote, per member, in both chambers.
clerk.house.gov roll-call XML; senate.gov roll-call XML
House records identify members by their official bioguide id; Senate records use LIS ids mapped through the public roster. Votes are matched on those identifiers.
What each bill is about
Each voted-on bill's policy area, and the Congressional Research Service's plain-English summary of what it does.
govinfo.gov bulk data: BILLSTATUS
Policy areas are assigned by the Library of Congress, one per bill, so we do not choose which bills appear. CRS summaries are quoted word for word.
Congressional roster
Who currently serves, in which chamber, and each member's Federal Election Commission identifier.
unitedstates/congress-legislators
Members are matched to filings by identifier, not by name. Candidates may file under any form of their name, so names do not match reliably.
Zip code to congressional district
Which district each zip code sits in. This powers the library's zip search.
US Census Bureau: 2020 ZCTA to 119th Congressional District relationship file
The Census maps its own approximation of each zip, so the match is close but not official. A zip that crosses a district line shows every district it touches. Zips that are only PO boxes have no Census area and return no match.
What we do not yet use
We do not publish named individual donors. We do not cover state-level campaign finance (governors and state legislatures): no federal source holds it, and the licensing terms of the main private source do not allow a public site to republish it. We do not use lobbying disclosures to choose bills, because most filers do not name the bills they lobby on.
Copyright of the underlying data
Federal Election Commission data is produced by the United States government and is in the public domain. Our writing, design, and analysis are our own.
How we handle these sources is described on the methodology page.