For data marketplaces
What a CorpusReady pack contains, and how you can receive submissions.
What a pack contains
| File | Contents |
|---|---|
| Clean dataset (CSV) | The de-identified data |
| Redaction log | Field, entity type, count and method |
| Provenance PDF | Summary, origin, date coverage, field dictionary, languages, de-identification method and k value, rights summary, licence basis and the business’s attestation |
| Manifest (JSON) | The same structure, machine-readable |
| Checksums | For the clean file |
| Scorecard | With an indicative price band |
Every document in the pack is built from metadata and the redacted sample only. Rights results are flags from a fixed questionnaire, not legal advice. Value scores are indicative, not an offer.
Manifest schema
The manifest follows a JSON Schema: corpusready-manifest.draft.json. It is a draft. Pack generation is not live yet, and the schema will be fixed and versioned when it is.
Receiving submissions
A pack is only ever sent when the business chooses to send it, and only to the marketplace it picks. By default, a submission arrives by email with the pack attached and a referral code. Where a marketplace offers an API, submissions can go through it instead once that integration is built.
No packs have been submitted to any marketplace yet. If you run a marketplace and want to set the address, minimum dataset size or format you would like to receive, email hello@corpusready.com.