From Glossary to Dataset
Using Queering Glamorgan’s historical vocabulary to search a century of Welsh newspapers
Digital archives, historical data and the web.
Using Queering Glamorgan’s historical vocabulary to search a century of Welsh newspapers
Testing a locally run pipeline that detects shots, transcribes speech, describes scenes and makes public-domain films searchable.
A self-hosted prototype for finding places in text, reverse geocoding coordinates and linking the results to Wikidata.
A history-hackathon experiment that gives archival portraits animated eyes and uses webcam tracking to follow the viewer.
An experiment using face-api.js and OpenSeadragon to zoom into an IIIF image as the viewer moves closer to the screen.
A hackathon workflow using Llama 3 and ElevenLabs to summarise historical newspaper articles and turn them into spoken audio.
A map-based visualisation following bus traffic across Wales over a full day.
Analysing the activity, sentiment and recurring language of a prolific commenter on the Cambrian News website.
An NLP workflow for linking names in historical newspapers and journals
A National Library of Wales hackathon experiment visualising faces in Welsh Newspapers Online over time, grouped by month.
Exploring crimes, criminals and punishments in the gaol files of the Court of Great Sessions in Wales from 1730 to 1830.
The People’s Collection Wales passed 90,000 uploads, making it a good moment to explore the data behind the website.
Recreating a visual timeline of newspaper front pages using the National Library of Wales’s historic newspaper collection.
Using machine learning to identify recipes in more than 200,000 articles from the Welsh Newspapers Online collection.
The National Library of Wales have made available transcriptions of shipping records from 1856 to 1914, which were transcribed by volunteers at the National Library.