Research data management · Web development
FactGrid2GND
Turning research data into national authority records
FactGrid accumulates person records at scale, but they remain inside their own Wikibase universe. For research results to stay citable and discoverable in library catalogues, they need to reach the Integrated Authority File (GND) of the German National Library. Until now, that path was manual work.
The task
Prepare person records from FactGrid so the German National Library can ingest them automatically. This demands an exact mapping from the Wikibase data model to MARC 21 under the library’s delivery specifications — including a check for whether an authority record for a given person already exists.
The approach
At its core sits a Python conversion layer that retrieves FactGrid statements via SPARQL and translates them into MARC 21 fields. A match against a local copy of the GND person data prevents duplicates; where the local database comes up empty, the tool queries lobid.org and caches the responses. For environments without network access, FactGrid itself can be built into a local SQLite database from a dump, so the entire conversion runs offline.
An Angular interface ties it together: switch the data source, run the
conversion, and inspect the result before the .mrcx file leaves the
institution.
The outcome
A research database becomes a supplier to the national authority infrastructure. What used to be transferred record by record is now a reproducible step in the workflow — and the link between project data and library catalogue persists.
A similar undertaking?
Tell me about it — you will get an honest assessment of effort and feasibility.
Get in touch