Le jeu de données Evidoria

Everything published on this site is one reusable dataset: evaluated practices with rubric-based quality scores, study-design evidence strength, structured narratives, funding-instrument tags, regions and full provenance. This page is its stable description — what's in it, how it's versioned, how to get it, and how to cite it.

Ce qu'il contient

Catalogue	Pratiques
AI in Education	266
AI in the Public Sector	267
City Innovation Library	399
Digital Inclusion	610
Gender Equality	720
PES & Nature-Based Solutions	371

Accès

• Téléchargement CSV — one row per practice; add any Browse filter as query parameters.
• REST API — practices, evaluations (criterion-level), catalogues, ontology; paginated JSON.
• DCAT-AP — machine-readable catalogue description for open-data harvesters.

Versionnage

The corpus changes nightly (new practices, enrichment, curator review), so the dataset is date-versioned: the version above is the date of the last change to any published record. Cite the version you used; the CSV you download reflects exactly that day's corpus.

Licence

Dataset descriptions and structured metadata are available under CC BY 4.0 — reuse with attribution to “Evidoria — the Evidence Observatory, by C-NAPSE”. Source documents linked from practices remain under their publishers' terms.

Comment citer

Méthode et fiabilité

Scores follow one documented rubric method (methodology), including how our evidence ladder maps to the Maryland SMS, Nesta and EMMIE frameworks. Every record carries a validation level; an inter-rater reliability study of the rubric is on the research roadmap and will be linked here when published.