Lightcast ↔ CCM
Employer entity × year × public-firm link, with time-varying CCM identifiers and explicit corporate-history decisions.
An accounting-first research infrastructure project for making employer-to-firm and employer-to-company linkage decisions inspectable, reproducible, and explicit about their evidence limits.
The hub separates two substantive linkage objects while applying one validation vocabulary. Lightcast↔CCM is first because it connects job-posting measures to public-firm accounting and market identifiers; Lightcast↔Revelio follows; the benchmark governs reliability claims for both.
Employer entity × year × public-firm link, with time-varying CCM identifiers and explicit corporate-history decisions.
Employer entity × year × company link, preserving company-versus-parent distinctions and temporal ambiguity.
Development and frozen-holdout separation, metric definitions, error taxonomy, and maturity gates shared across both linkages.
Each resource page defines the source and target entities, matching unit, intended coverage, public disclosure boundary, construction stages, available evidence, failure modes, and accepted limitations.
A planned annual employer-to-public-firm linkage that connects job-posting company entities to time-varying accounting and market identifiers.
A planned company-year crosswalk for reconciling employer entities across job-posting and employee-history data while preserving parent, subsidiary, and temporal ambiguity.
The benchmark page distinguishes development/adjudication evidence from a frozen holdout and from a validated release. No model score is presented as ground truth, and no partial candidate-recall estimate is labeled overall recall.
This prototype specifies research judgments that a future production linkage must expose rather than compress into a final merged dataset: entity eligibility, candidate generation, firm-versus-security identity, organizational level, temporal validity, unresolved cases, and the labeled evidence required for any reliability statement.
Defines how a future versioned linkage must separate source decisions, candidate rows, accepted links, temporal intervals, and downstream identifiers.
Separates rule development from frozen-holdout evaluation and reports unresolved cases rather than optimizing only for match rate.
Publishes methods, aggregate diagnostics, provenance, and synthetic examples while keeping licensed records and private identifiers outside the public build.
The package is built from an explicit allowlist. It contains only original public-safe documentation, generic code, and synthetic examples. The release manifest records checksums for every downloadable object.
Measurement scope, matching stages, evidence ceiling, validation gates, and disclosure boundary.
Download MethodologyReserved synthetic identifiers and .example domains for testing the reference implementation.
Download Synthetic exampleGeneric input and candidate-output fields without licensed values.
Download Public schemaSelf-contained TypeScript source, package metadata, tests, license, and synthetic inputs; extract it, run npm install, then npm test.
Download Reference implementationAllowlisted files, byte sizes, media types, and SHA-256 checksums.
Download Release manifestMachine-readable citation metadata for the prototype method release.
Download Citation fileLicense for original public-safe code and artifacts only; source-dataset rights are excluded.
Download LicenseVersion history and explicit statements of what each release does not contain.
Download ChangelogYang, Hongye Chris. 2026. Open Accounting–Labor Data Linkage Hub: Public-Safe Entity-Linkage Methods and Validation Protocol, version 0.1.0.
A more favorable label cannot be earned by visual polish or code completion. It requires the benchmark, provenance, reconciliation, and verification evidence defined for that release stage.
Methods and public-safe artifacts are inspectable, but production linkage evidence is incomplete or not released.
The resource is usable end to end, with visible benchmark evidence and limitations, but may change materially.
The release has passed its frozen holdout, predeclared intended-use performance criteria, reconciliation, provenance, and independent-verification gates.
The canonical private linkage sources, adjudicated development sets, frozen holdouts, and benchmark-supported production metrics are not public evidence in version 0.1.0.