Methodology
Conflict Score (0–99)
For each member of Congress we measure the gap between who funds them and how they vote. For every industry that donates to a member, we compute an alignment ratio — the share of that member's votes on industry-relevant bills that went the industry's way — and weight it by donation size: alignment_ratio × log10(donation_cents + 1). The log keeps a $5M donor from drowning out everything else. Per-industry terms are summed, then percentile-normalized across the chamber to 0–99. Higher = votes track donors more closely.
Bill-to-industry relevance is determined by topic classification of each bill against a 13-category industry taxonomy (Finance, Energy, Health, Defense, Technology, Real Estate, Agriculture, Transportation, Communications, Labor, Law/Lobbying, Retail, Education).
Match Score
Your Political Fingerprint is built from your reactions (agree / disagree / skip) to real bills. A legislator's Match Score compares your stated positions to their actual roll-call votes on the same bills, weighted by the topic importance you set. We require at least 5 reactions before showing a score, because fewer makes the number noise.
The six stats (PWR, CLT, BIP, EFF, TRM, LOY)
PWR — committee power: committee weight × leadership-role multiplier. CLT — clutch: bills passed in end-of-session crunch / bills introduced. BIP — bipartisanship: cross-party cosponsorships / total. EFF — effectiveness: bills enacted (full credit) + bills cleared committee (quarter credit) / bills introduced. TRM — seniority: 20 points for the first term, +10 per additional. LOY — party loyalty: party-line votes / total votes. All normalized 0–99 and percentile-ranked within the chamber.
Data sources and update cadence
Votes, bills, and committees from Congress.gov; campaign finance from the FEC and OpenSecrets bulk data; ideology coordinates from VoteView (DW-NOMINATE); district resolution via the US Census Geocoder (we never store raw addresses, only the resolved district like "CA-11"). The pipeline runs nightly at 02:00 UTC.
Honesty notes
Correlation is not causation: a high Conflict Score shows votes that track donor interests, not a proven quid pro quo. Coverage is uneven — donor-linked vote coverage is more complete in the House for the 119th Congress. AI-generated narratives (report cards, summaries) are grounded in the structured data above and labeled as AI-written.