Data Sources

Data Sources shows which inputs were used, where they came from, and whether they satisfied the declared data contract.

Per-series fields

• canonical series name;
• vendor and venue;
• available date range;
• sampling frequency;
• point-in-time and survivorship properties;
• synthetic or fallback status;
• freshness;
• immutable fingerprint or Merkle root where available.

Compatibility

Series used together must be semantically compatible. Funding from one venue must not silently drive returns from another venue. Daily and intraday series cannot be substituted without explicit aggregation and timing rules.

Missing data

Unavailable inputs become explicit requests. The view should distinguish a genuinely missing series from an unsupported claim semantic, a licensed dataset, or a parser artifact.

Public display

A public site may show provenance and fingerprints without redistributing licensed values. Always review the source's redistribution terms.

Machine records remain immutable. Human-readable source titles and summaries live in the presentation plane: enable ingest titling with PENROSE_SOURCE_TITLER=1, use the presentation overlay for existing records, and run make lint-ui so coded fields do not reach the dashboard without a readable form.