Visualising Annotations - NLW Transcription projects Glen Robson - - PowerPoint PPT Presentation

Visualising Annotations - NLW Transcription projects Glen Robson - IIIF Technical Coordinator Image from: http://map.coflein.gov.uk/index.php?action=do_images&cache_name=&numlink=23303#tab

PROJECTS • Aberystwyth Student records • From 1870 to 1910 • 8 Volumes • Partnership between NLW and Aberystwyth University • Transcription being done by Alumni in Cardiff, Aber and other places

PROCESSING STAGES 1. Map annotation body to Linked Data fields 2. Data cleanup 3. Reconcile 4. Load to SPARQL DB 5. Repeat from 2.

Admission Date from dateutil.parser import parse date = parse(value, fuzzy=True)

DATA ERRORS • 2 Types: • Transcription errors • Source data errors (or lack of consistency)

DATE ERRORS • Out of 378 pages / people • Only 3 invalid dates

Reconciliation with WikiData • Mixed results • UK Schools have changed a lot since 1870! • No longer have Grammar schools • Many schools don’t have Wikidata (or wikipedia) information • Advantage of Wikidata is you can add them.

SUMMARY • Annotations nice to work with • Can check results against an image • Process them as json or LinkedData • From data • Contains lots of useful data • Historical type of database • Reconciliation not easy…