Data Sources & Methodology

Source: NYS Open Legislation API — official New York State Senate developer API.

Coverage: Floor session transcripts (2023–present) and committee hearing transcripts (2023–present). Full verbatim text as published by the Senate.

What is citable: The verbatim transcript text, roll call votes, bill references, and speaker attributions are direct API records. Cite as: New York State Senate [Committee], [Date], NYS Open Legislation API record #[ID].

What is AI-generated: Debate summaries, wire briefs, testimony abstracts, and senator engagement summaries are Claude-generated from the transcripts. Label these as AI summaries if referenced.

Updated: continuously via API

Source: NYS Board of Elections — certified general, primary, and special election results.

Coverage: 1996–2025. State Senate, Assembly, Congress, U.S. Senate, statewide offices, and Presidential races. NY fusion voting: votes are summed across all party lines per candidate.

What is citable: All vote totals, margins, and winner/loser data are certified NYSBOE records. Margins are computed from certified totals.

Updated: per election cycle

Source: NYC Board of Elections — Cast Vote Records (CVRs), the full anonymized ballot-level data released after each RCV election.

Coverage: 2025 NYC Primary — 38 contests, 4,321,937 ballots. Includes every rank recorded on every ballot.

What is citable: Round-by-round vote totals, elimination order, vote transfers, exhaustion counts, and ballot depth distributions are all computed directly from the CVR source files.

Note: RCV analytics (mandate gap, transfer loyalty rates, borough comparisons) are derived calculations from the CVR data — they are not published by NYC BOE.

Loaded from CVR files released July 2025

Source: U.S. Census Bureau, American Community Survey — 2024-year estimates (covering 2020–2024).

Coverage: All 63 NYS Senate districts. Population, income, poverty, race/ethnicity, education, health insurance, employment, veteran status, disability, commute mode, foreign-born, limited English households.

What is citable: All figures are published ACS estimates. Cite as: U.S. Census Bureau, American Community Survey 2024-Year Estimates, Table [variable code], [district].

Note: ACS estimates carry a margin of error not displayed on this platform. For precision analysis, consult the Census API directly.

ACS 2024-year estimates (released Dec. 2025)

Source: NYS Board of Elections — monthly voter enrollment by party and Senate district.

Coverage: Registered Democrats, Republicans, Independents, and other parties per district.

What is citable: Party registration totals as published by NYSBOE. Cite the enrollment file date.

Last loaded: 2026-03-23

Source: NYS Board of Elections via data.ny.gov — itemized monetary contributions to campaign committees.

Coverage: All 63 active NYS Senate campaign committees (2022–present). Includes donor name, amount, employer, and contribution date.

What is citable: Individual contribution records are official NYSBOE filings. Totals, donor rankings, and industry classifications are computed from those records.

Limitation: Committee filer IDs are maintained manually and may not reflect newly registered or alias committees.

Data through: 2026-03-28

Source: NY Commission on Ethics and Lobbying in Government via data.ny.gov — bi-monthly lobbying disclosure filings.

Coverage: Lobbying contacts targeting NYS senators, by client organization and subject area.

What is citable: Contact counts and subject areas are from official NYCELIG disclosure records.

Limitation: Name matching uses automated normalization — approximately 7 of 63 senators may not match due to accented characters or unusual name formatting, and would show zero contacts despite actual lobbying activity.

Updated: annually from NYCELIG filings

Method: A scenario model computed from voter registration (40% weight) and recent contested general election margins (60% weight), using up to the last 4 elections with margins under 40 points.

Scenario shifts: ±5 percentage points applied to the base lean to model favorable/unfavorable national environments. Generic ballot baseline from Silver Bulletin (Nate Silver). Current: D+5.7, as of 4/19/2026.

What is citable: The underlying registration and election data are citable. The scenario ratings (Safe/Likely/Lean/Toss-up) are a derived model and should not be cited as authoritative election forecasts.

Model inputs: NYSBOE enrollment + certified election results

Source: Silver Bulletin (Nate Silver) — daily generic congressional ballot polling average.

Coverage: National D/R margin from aggregated polling. Used on the district map and in vulnerability scenario modeling.

Limitation: Fetched hourly and cached. If the fetch fails, the last cached value is shown without an age warning. Check the date stamp on the district map for freshness.

Current: D+5.7 as of 4/19/2026

Source: NYS Senate Leadership Roster and Senate member pages.

Coverage: Leadership titles (Majority Leader, committee chairs, etc.) and party affiliation for all 63 senators.

Limitation: These are maintained as a manually updated list in the platform configuration. They are not fetched from an API and may lag after leadership elections or mid-session changes. The Open Legislation API does not publish party or leadership data.

Last updated: 2025-2026 session (manually verified)

Method: Generated by Claude (Anthropic) from the senator's floor speeches, committee hearing testimony, and sponsored legislation. Topic tags are sized by number of citations; each links to the source bill or hearing record.

What is citable: The linked bills and hearings are citable. The topic labels and summary prose are AI interpretations and should not be cited as the senator's stated positions.

Known limitation: A senator who asked a neutral clarifying question at a hearing may be coded as engaged on that topic. Inferred topics are not validated against any authoritative source.

Regenerated periodically from new transcript ingestion

Method: Wire briefs, topic summaries, and testimony abstracts are generated by Claude (Anthropic) from verbatim hearing transcripts obtained via the Open Legislation API.

What is citable: The verbatim transcript (linked from each hearing page as "View full transcript → Archive") is the authoritative record and is directly citable. Summaries are editorial interpretations.

Testimony claims: Key claims extracted from testimony are flagged as "verifiable" or not by the AI — this is not an independent fact-check, it is the AI's assessment of whether the claim is empirically checkable.

Generated at time of transcript ingestion