Addressing the growing need for innovation and automation in library cataloguing systems, this cutting-edge book teaches readers how to build AI-powered cataloguing apps using Microsoft Power Apps.
This authoritative guide provides a step-by-step approach to how cataloguing processes can be modernised by using artificial intelligence, featuring a major retrocataloguing project as the primary case study. Focusing on creating inclusive bibliographic records, this book will enable readers to train AI models, develop tools to improve efficiency and collection management and address challenges such as uncatalogued collections and the demand for more accessible, enriched metadata. Chapters
An introduction to cataloguing and metadata Getting started with the Power Platform ecosystem Integrating artificial intelligence Building, deploying and maintaining the cataloguing app With a focus on practical applications and user-friendly methodologies, this transformative book provides librarians, information professionals and data managers with an accessible guide to building their own automated cataloguing systems.
Page 37 – This book mainly seems to be for libraries that don’t have something like Horizon or OCLC. This Dataverse seems to be the equivalent of Horizon, only you have to make authority connections manually where Horizon makes them automatically. So, it’s more work. Then the Power Pages from a couple of pages ago are for libraries that don’t have a public-facing catalog page. The library that did this still had a card catalog.
Page 64 = About automating subject headings. This also isn’t feasible because there is a limit of 200 tags. There is a workaround by just creating multiple sets, but it isn’t practical considering there are 388,594 subject headings as of April 2025 (See https://www.loc.gov/aba/publications/...). You’d have to create 1,942.97 sets. Even then, you’d still have to do the headings manually because this is just for the main heading, because it doesn’t seem to have the capability of putting headings together, like the $x $v or $y subdivisions. There is also no mention of how free-floating subdivisions would work.
To be honest I didn't finish this book. By page 100 I had determined that the solutions were impractical for my library, and also needing to build your own subject headings suggester is already unnecessary as of December 8th, when OCLC introduced their AI subject headings tool.
This means that I stopped right before the book got into the nitty-gritty of actually building the application and maintaining it. It might be an excellent book as far as that goes, but since the whole thing seems unnecessary, it's a bit of a moot point.