Datasets

Menowise aggregates journal-evidenced information on menopause and midlife women’s health. Every dataset is a text file in the repo (data/), validated in CI against a JSON Schema (schema/), and served here verbatim.

Not medical advice. These datasets aggregate and organize what the published literature reports, with citations and honest evidence-strength labels. They do not diagnose, advise, or recommend treatment.

dataset	files	schema
signals (per-source claims)	`data/signals/*.yml` (empty; seeding pass)	`signals.schema.json`
entries (canonical, cited)	`data/sites/*.yml` (empty; seeding pass)	`sites.schema.json`
entities (controlled vocabulary)	`entities.yml`	`entities.schema.json`
sources (fetch allowlist)	`sources.yml`	`sources.schema.json`

The model is the info archetype’s, unchanged in shape: per-source signals → canonical entries with provenance and confidence (kdc is build 1). Field detail: founding/schema-sketch.md.

License

The datasets published by this project are licensed under CC BY 4.0 (studio default, ADR-0002). The project’s code is MIT-licensed (see the repo’s LICENSE).

Provenance

Where the data comes from: peer-reviewed biomedical literature and reputable clinical guidelines, registered in sources.yml (the fetch allowlist, ADR-0025). Every published entry cites its sources.
How it is collected: gn_info_scout extracts per-source signals (one claim per source) from allowlisted literature; gn_info_records clusters signals into canonical entries with an honest evidence_strength label. Both are studio agents; every change arrives as a human-merged PR.
Transformations: extraction and clustering only — no averaging of conflicting claims, no inference beyond what sources state. Null and negative results are recorded, not dropped.
Current state: the datasets are empty pending the seeding pass (careful, cited entries; menowise#4). Honest-zero entries (surveyed-thin/surveyed-empty) are content, not gaps (GD-0004).