Present:

Apologies:

Agenda:

This is a communications focused agenda.

  1. Investigating gendered language in the archive - Tasha Kitcher

Proposal document

Proposal document

  1. Transforming personal researcher notes into archival descriptions - Tasha Kitcher

Proposal document

Proposal document

  1. Linking communications collections through computer vision - Tasha Kitcher

Proposal document

Proposal document

  1. Creating a searchable database of General Post Office circulars - Tasha Kitcher

Proposal document

Proposal document

  1. Textiles machinery Wikidata game (connected to both Deeper catalogue data & Object name vocabularies) - Tim Boon

Notes:

  1. Comparison between Generize.io and the Heavens NLP tool. The NLP tool - 8 annotators trained the NLP using 10% of the collection, then ran on the rest of the collection. Want to understand, if gender is problematically represented, is that followed through into the secondary literature, or is it in the archival material itself (rather than just the cataloguing). Authority records pretty well developed. Q from Stefania about how this might help with connectivity. Tasha - this is being developed more towards something that is done before connectivity, it is more about understanding positionality. Number of dependencies - waiting for Lucy to open her GitHub to the project and we’ll need to see that before we know how easily we can use it. Come to the investigation meeting again if the model isn’t in a useable format. Q from Alex - do you have the technical skills to do this as you are? Tasha - we think we have enough to have a go! University of Newcastle are also testing the tool. Agreed we should contact Uni Newcastle too. Q from Kaspar - how will you evaluate whether Lucy’s model works. Tasha - would need to do own versions of annotations and do comparative work there. Worth Tasha and Anna-Maria discussing and defining the approach in more detail. Comment from Tim - understanding gender is a good way to start. Be good to think through ethnicity, class and age too. Proposal approved to go forward - the first steps of it and then come back to us with any initial insights.
  2. Look at research notes and understand how they can be transformed into catalogue descriptions. Do a small scale proof of concept with the Telephone heritage group and then decide whether to expand into digitising Bradford related material? Is it worth thinking of the Ipsophone as worth digitising anyway? Even if this investigation doesn’t have a Bradford strand. Stefania - it would be good to think about how audio notes can be incorporated too, based on learning from the folksongs investigation with Jen Reid. Proposal approved for a small proof of concept version, then come back to the investigations meeting.
  3. Want to bring together NMS and SMG collections based on how they look. Use the Heritage Weaver process developed in Kaspar’s work and think through what the visual connections actually are. What kind of research can actually be done through connecting materially based on how they look. This investigation does linkage first and develops research questions based on that. This will work to develop Kaspar’s investigation too, could potentially help with the multimodal tagging of collections too. How quickly can we access the NMS image data, and ingest it? 2 weeks from request? SMG collection potentially doesn’t have many linked photos. Tim, Alex, Arran to bring up with Adrian Hines next week. Proposal approved.
  4. This would take up the most amount of time of all the investigations. Build on Asa’s GPT work. Circulars all digitised. Exchange book isn’t digitised - but could do so, just focusing on the Bradford pages. OCR work to happen, whilst Asa looks into how the GPT work might be applicable to this. Tasha and Alex to meet with Asa. Proposal approved. Keep an eye on how much work this is requiring, in conversation with Asa.