Work from the official record, then show the method.
GovData organizes available people, bodies, meetings, matters, motions, vote records, and source references into a source-linked public-record graph. It can support discovery and analysis, but the graph does not supply an official's intent, a ready-made score, or a complete person-level record.
Corpus scale needs context.
Rounded current corpus floors include 189,000+ elected officials and 21M+ recorded vote records. Not all corpus vote records are per-person roll calls. Person-level roll-call linkage, attendance, sponsorship, identities, terms, historical depth, source links, and coverage vary.
The records support research, not automatic conclusions
Scorecards, coalition analysis, reporting, research, and apps are user-built workflows. Each requires a stated methodology, source review, identity checks, treatment of missing data, and a correction process.
Research a specific public-record question
Reporters can search for available votes, matters, sponsors, terms, and source documents, then verify the relevant record before publication.
- Define the official, body, issue, and period
- Distinguish roll calls from aggregate actions
- Review source text and attachments
- Disclose missing or uncertain records
Publish the choices behind the score
A user can classify issues and calculate a score only after defining included motions, weights, absences, abstentions, denominators, and update rules.
- User-defined issue taxonomy
- Documented inclusion and exclusion rules
- Source review for classified motions
- Visible methodology and correction policy
Analyze a reviewed roll-call subset
Voting-similarity or coalition analysis depends on which bodies, periods, matters, vote values, and missing records the user includes. Similarity does not establish motive or affiliation.
- Reviewed person-level roll calls
- Explicit similarity method
- Identity and membership checks
- No inference of private intent
Build a documented analytical dataset
Researchers can define a sample and method, assess source and jurisdiction bias, review identity matches, and preserve the data vintage used in the study.
- Sampling and missingness review
- Field-level coverage assessment
- Versioned extracts and methods
- Reproducible limitations
Design for uncertainty and correction
Civic products can use free public search, records retrieved through paid API access with an issued key, or separately scoped extracts, but the application must explain scope, dates, classifications, source links, and known identity limits.
- User-facing source context
- Coverage and freshness notices
- Identity review and correction path
- No unsupported candidate profile
Use source records with neutral caveats
Users may research an incumbent's available public record, but they must define the issue framework and verify the source before making evaluative or campaign claims.
- Selected public votes and matters
- User-owned issue classifications
- Disclosed gaps and date limits
- No claim of a complete candidate record
Several record types, with different linkage depth
The graph preserves available entities and relationships. It does not imply that each person, matter, motion, vote, sponsor, term, attendance entry, document, or source link is present.
Officials, seats, bodies, and terms
Available names and membership records can connect a person to a body or term. Appointment, election, seat, title, and end-date detail varies by source.
Items, actions, and available sponsorship
Agenda matters and motions may include titles, text, status, sponsors, attachments, and history. Field availability and meaning differ across systems and jurisdictions.
Aggregate actions and person-level roll calls
A recorded vote record may be an action outcome, tally, motion-level entry, or person-level choice. The 21M+ corpus floor includes these record forms and is not a count of 21M individual officials' roll-call choices.
Links and documents where available
Meeting pages, agenda records, minutes, packets, attachments, and source URLs may support review. Links can move, documents can be missing, and source text can change.
Names and labels require review
A stable data key can support analysis, but it is not a guarantee that the underlying person identity or classification is correct.
Identity-resolution risks
- Two people with the same name can be merged incorrectly
- Name variants can split one person into multiple records
- Seat changes, body changes, and nonconsecutive terms can confuse linkage
- Titles, suffixes, appointments, and spelling can differ across sources
Classification limits
- Topic tags and issue categories may be research classifications
- Source-native fields and derived fields must remain distinguishable
- Ambiguous motions can require document-level review
- Users must document classification, weighting, and correction rules
GovData classifications are not official determinations, allegations, endorsements, or complete candidate records. Same-name and identity-resolution errors are possible. Consequential person-level claims require review of the original source, the entity match, the official's service context, and the user's methodology.
Corpus scale is not field-level completeness
Person-level roll-call linkage, attendance, sponsorship, identities, terms, historical depth, source links, and coverage vary by jurisdiction, body, time period, system, and record type.
Availability is not uniform
Indexed jurisdictions may expose different bodies, record types, date ranges, and fields. A corpus total does not establish coverage for a specific council, board, authority, or period.
Start and end dates differ
Some records reach further back than others. Portal migrations, archival practices, corrections, and late publication can create gaps or change the available history.
Presence is source-dependent
Attendance, sponsorship, cosponsorship, excused absence, recusal, and participation fields are not published or encoded consistently. Missing data should not be treated as a negative fact.
Links support checking, not certainty
Source links and documents are included where available. Users should verify the current official record, amendments, corrections, minutes, and operative documents before publication or analysis.
Absence from a search result does not prove that an official, term, meeting, motion, vote, sponsor, attendance record, or source document does not exist. Confirm the intended field coverage and historical window before drawing a person-level or comparative conclusion.
Start with free public search or paid API access
Current access includes free public LocalGovIndex search. API retrieval is paid and requires an issued key. Extracts and research samples require separate scoping for people, bodies, issues, dates, fields, identity review, source review, and delivery terms.
Available starting points
- Public LocalGovIndex search and record pages
- Paid API search and retrieval with an issued key under applicable terms
- Schema review and representative records
- Separately scoped official, body, or issue research samples
Requires separate scoping
- Person, body, jurisdiction, period, or issue extracts
- Identity-resolution review and alias decisions
- User-defined classification and methodology fields
- Refresh schedules, corrections, and application integration
This page does not promise universal person-level linkage, complete voting or candidate histories, perfect identity resolution, continuous attendance or sponsorship data, a specific update cadence, or durable third-party source links. Confirm scope and delivery terms in writing.
From discovery to a verified voting denominator
The Chicago case audit starts with a complete source-defined frame, checks person-level statuses against the Clerk's grouped roll-call report, and publishes the method and limitations.
Scope an accountability research sample
Bring a person, body, issue, jurisdiction, period, and intended workflow. We can test the available fields and sources, surface identity and coverage gaps, and define a realistic sample.
A useful first review
- Define people, aliases, seats, bodies, jurisdictions, and dates
- Specify roll-call, motion, sponsorship, attendance, or term fields
- Review representative results and available source links
- Document identity, coverage, history, and classification limits
- Agree on the user's methodology, corrections, and delivery scope