About Officium
How every number on this site is made — and what it does and doesn't mean.
What Officium is
Officium gathers the public paperwork of the U.S. Congress — stock trades, financial disclosures, lobbying and foreign-agent registrations, campaign money, travel, ethics cases, and votes — into one profile per member, so journalists, researchers, and ordinary citizens can see potential conflicts of interest without digging through a dozen government websites.
Where every number comes from
Every figure on this site traces back to an official public record. We do not invent data; we ingest it, index it, and label where it came from.
- Stock trades — the stock-trade disclosures (PTRs) members must file under the STOCK Act with the House Clerk and the Senate, ingested via QuiverQuant's PTR data. The trades page shows the data's own “as of” fetch date. For context: hundreds of current and former members have trade filings on record going back to 2012 — most members don't trade.
- Annual financial disclosures (PFDs) — the yearly reports of assets, debts, outside jobs, and gifts that House and Senate members file, parsed from the House Clerk and Senate eFD systems, with links to the official PDFs.
- Foreign-agent registrations (FARA) — who represents foreign governments and interests in Washington, from the Justice Department's daily bulk files.
- Lobbying registrations (LDA) — who lobbies, for which clients, on which issues, from Senate and House lobbying-disclosure filings, including former members who became lobbyists.
- Campaign money — receipts, donors, and outside spending from the Federal Election Commission's official bulk downloads, including the 2026-cycle candidate files (cn26 / weball26).
- Votes — roll-call votes from GovTrack (a recent sample per member, not full voting history).
- Election odds — where shown, market prices come from Polymarket and are attributed as such. They are betting-market prices, not our predictions.
How research flags work
Our software compares a member's money — stock trades, donors, disclosed assets — against their official work, like votes, bills, and committee hearings, and marks the overlaps. Flags are leads to check, not accusations. Some flags are about timing (a stock trade days from action on a related bill); others are about a direct connection with no dates involved at all, such as a hearing witness who also gave the member money.
- Strong / Moderate / Weak is match strength, not seriousness. It says how tightly the pattern our software looks for was met: for timing flags, how close the dates fell; for the rest, how direct the connection is. A strong match is a computer's measure of a pattern, never a judgment about anyone's conduct — and the strongest flags are simply the ones worth checking first.
- Some flags have no source link. Roughly 18% of flags could not be tied back to a specific record we can cite, and are labeled unverified. Treat those with extra care.
- A flag is never proof. Members trade through advisers, in blind trusts, and by coincidence. Always check the original filings before repeating a flag as fact.
What we don't do
- No invented dollars. When a source publishes ranges (as disclosure forms do) we show ranges or clearly labeled estimates — we never fabricate a precise figure the record doesn't contain.
- No claims about intent. We report what was filed and when. We do not claim to know why anyone traded, voted, or accepted a trip.
- “No candidacy on file” means exactly that. It says no filing exists with the FEC yet — it is a statement about paperwork, not about anyone's plans or intentions.
Freshness
Government data moves at different speeds, so every dataset carries its own “as of” date, shown next to each file on the data catalog. If a number matters to you, check its date first.
Repairs & honesty
- We fix parsing problems in the open. Example: a repair to the financial-disclosure (PFD) liability column re-parsed misread rows, every dollar amount is validated per row (a corrupted range can never enter a total), net worth shows as unknown — never $0 — when a filing could not be parsed, and rows we could not resolve are explicitly flagged rather than silently dropped or guessed.
- Our gaps are published as data. Coverage is incomplete for many members, and we ship the audit that says exactly what is missing — browse it on the gaps explorer or download the raw file. A thin profile usually means we haven't ingested a source yet, not that nothing happened.
- Official records and research flags never mix. Trade tables contain only real filings; research flags and badges are always labeled as computed by software. See the glossary for plain-English definitions.
Everything is open
There are no accounts, no paywalls, and no locked tiers. Every page — including full history and former members — is open to everyone, and every dataset is downloadable as bulk JSON from the data catalog (static files under /data/*.json). The site is deliberately crawlable: a full sitemap, structured data on profiles, and one-click CSV/JSON exports on key views.
Built with Next.js, TanStack Query, Zustand, and D3. Government records carry no copyright; the stock-trade compilation (QuiverQuant) and the roster and vote data (GovTrack) are third-party sources with their own terms and are not ours to relicense. Data is public domain / CC0 where applicable. This site exists to make congressional influence research faster, deeper, and more accessible than the fragmented official sources or investor-focused alternatives.