Data

Data for Researchers & Developers

PartyLine's scores for members of Congress, with key roll call votes behind them, are published as JSON files that anyone can download. There is no key, no sign-up and no query endpoint. District-level opinion data is not published.

Politician Scores

Scores from 0 to 100 on 13 policy areas, for more than 2,000 members of Congress, past and present. Every current member has a record. 0 and 100 are the two sides of an issue, and a score near 50 is a mixed record. A member with too few counted votes on an issue has no score on it.

District Aggregates

Aggregate public opinion data for congressional districts — where voters in a district actually stand on each issue. This dataset unlocks per district once 100 verified responses exist there; no district has reached that threshold yet, so there is currently nothing to serve. User-identifiable fields are removed after 90 days.

Roll Call Source Data

For each member and issue, up to 10 key votes: the bill, the date, the chamber and roll call number, how the member voted, and which side of the issue that vote counted toward. The files give the number of votes behind each score; they do not list every one of them.

Files You Can Fetch

These are the files this site's own pages load: plain JSON at fixed addresses under https://www.mypartyline.org. A request needs no key, no account and no special header. You download a file and read it; nothing answers a query.

  • /output_v2/index.json: every member in one file, about 2 MB. Each row of politicians has a bioguide_id, name, party, state, chamber, is_current, total_votes and scores, the member's scores by issue. last_updated is when the file was built.
  • /output_v2/issues.json: the 13 issues. For each one, description_a is the side a score of 0 stands for and description_b is the side a score of 100 stands for.
  • /output_v2/politicians/{bioguide_id}.json: one member's record, about 200 KB; put a bioguide_id from the index in the address. Under issue_scores, each issue listed has its score, a label in words, effective_votes (the votes that counted) and key_votes, up to 10 votes with date, bill, chamber, rollnumber, vote and direction.

The scale. Every score runs from 0 to 100. It is the weighted share of a member's counted votes on the issue that went to the description_b side (weighted_direction_b ÷ weighted_total × 100): 0 means every counted vote went to the first side, 100 means every one went to the second. A score is null where a member has fewer than 3 counted votes on the issue. A record can list fewer than 13 issues: for a member seated recently, an issue is left out of both files until the member's first vote on it. Read a missing issue as no score. The survey works in positions from −2 to +2, and a position sits on the same 0 to 100 scale at position × 25 + 50, so −2 is 0, 0 is 50 and +2 is 100.

Using the data. Reading these files needs no permission. Our Terms of Service (section 9) allow personal, non-commercial use; to republish the data or build a product on it, ask first at support@mypartyline.org. Please fetch at a pace that does not slow the site for its readers (section 10). The Methodology page explains how a vote is assigned to an issue and a side, and how votes are weighted.

Who Uses This

  • Academic researchers — studying voting patterns, political polarization, or constituent-representative alignment
  • Journalists & newsrooms — building issue trackers or interactive voter guides
  • Civic tech developers — apps that help voters research candidates or track how their representatives vote
  • Nonprofits & advocacy groups — analyzing voting records on specific policy areas

District-Level Voter Data

Are you a candidate, incumbent, or organization interested in district-level voter data? Tell us about your use case and we'll be in touch.