Vaucluse
Objective
Build a reproducible pipeline to extract, clean, and georeference postcards.
Current state & Next steps
- Industrialized by querying paid external servers (billing based on the volume of data processed).
- Workshop run as part of the distam+ summer school, on slides described in French and Farsi #academicvalidation
- Strengthen metadata cleaning with a database?
- Offer the service to other archival collections (and get paid).
- Consider additional sources?
- Industrialize outreach.
Progress log
- 26/02/2026: created the project page and rebuilt timeline history.
- 15/04/2026: the map for the Vaucluse archives is about to be published. HuggingFace portfolio space built. Offer put together for other archival collections. Benchmark done.
- 05/2026: first contacts with institutions. Considering (with Maxime) connecting to ongoing / in-development projects.
- 06/2026: organized a workshop day at the distam+ summer school “From Closure to Sharing: Exploring Research Data on the Heritage of Afghanistan and Neighbouring Territories”. Produced teaching notebooks and deployed our workflow on the IFRI corpus.
Comments