Research reproducibility

Bulk downloads

The source inventory and connector registry as bulk data — where the index looks, in machine-readable form, so a reader can check that it draws on what it says it draws on.

3Data products
39,965Total bytes
2Formats
deterministicGeneratedEvery product derives from package data, never from a clock.

Each product below is generated deterministically from the packages themselves rather than hand-maintained, so a download always matches what the pages and endpoints compute. Both describe where the index looks: which public records, which documented APIs, under which quotas and credential requirements.

The index’s own structure is not published in bulk. The outcome codebook as a machine-readable file, the JSON Schemas, and the relational DDL are together a data model ready to be populated — which is a starting kit rather than an audit trail — and the illustrative corpora are fictional data in that same real shape. Those remain available to operators through gii export.

3 products, served by @gii/api at GET /api/v1/downloads/{name} or via gii export <name>.

ProductTypeSizeContents
source-profiles.jsonapplication/json24.0 KBThe source inventory: every profile with its access interface, revision behavior, and what it feeds.
source-profiles.csvtext/csv4.1 KBThe source inventory flattened to CSV: id, publisher, tier, role, retrieval, auth, origins, domains.
connector-specs.jsonapplication/json10.9 KBThe connector registry: endpoints, pagination, credential env-var names. No credential values.
Illustrative reference data. GII-US is pre-pilot; no product contains a score or series for any real government decision. The methodology publications themselves are on the methodology page, and the API contract at openapi.yaml.