PROVENANCE + GOVERNANCE
Every data point should have a story.
Know where data came from, how it was collected, what rights govern its use, which transformations were applied, and how quality was validated.
Discuss provenance requirements- source
- linked
- permissions
- recorded
- transforms
- versioned
- qa
- accepted
LINEAGE INFRASTRUCTURE
Provenance travels with the data.
A clear, durable record makes each asset more legible to model teams, governance functions, legal reviewers, and technical operators.
Source
Origin, source type, collector, collection method, location, and timestamp.
Permission status
Participant, source, and program permissions linked to the relevant asset or collection.
Rights
Documented approved uses, scope, access boundaries, restrictions, and licensing context.
Transformation history
Versioned record of processing, derivation, formatting, alignment, and enrichment.
Annotation lineage
Schema version, annotation history, reviewer linkage, and semantic QA status.
Quality status
Technical and semantic acceptance criteria, review outcome, and data release state.
CONCEPTUAL RECORD
A machine-readable history of the asset.
The exact schema is tailored to the program, buyer requirements, data modality, and approved use. This example illustrates the structure only.
{
"dataset_id": "C360-001",
"source_type": "expert_capture",
"collection_method": "first_person_video",
"rights_status": "commercial_ai_training_permitted",
"expert_verified": true,
"annotations": [
"task_phase", "action", "tool", "outcome"
],
"qa_status": "accepted"
}PROGRAM-SPECIFIC GOVERNANCE
Controls should match the data and its intended use.
Governance is defined per program. Context360 does not rely on one broad claim to describe every source, modality, or operating environment.
Purpose-specific
Collection design, permissions, rights, access, and retention are aligned to the defined use case.
Traceable
Assets retain durable links to source, method, transformations, schema versions, and QA decisions.
Access-controlled
Non-public data is shared through controlled systems appropriate to the licensing and program requirements.
Reviewable
Records are structured so technical, operational, governance, and legal teams can evaluate the data lineage.