Illustration: Jean Jullien

Digital

Mia Ridge, Issue 119/02, 06.02.2019
Making digital collections more accessible
We’ve seen great progress in labelling images, transcribing audio and identifying people, places and concepts in texts with machine learning tools. The potential of combining these technologies with traditional cataloguing processes, to make digitised collections more accessible, is exciting. 

So when the British Library started talking to the Alan Turing Institute about how we could combine its expertise in data science and artificial intelligence with the library’s experience with digital scholarship, I had to be involved.

As a museum technologist-turned-digital curator, I know there’s a lot of interest in applying new technologies to collections data. I hope this project’s outputs – the code libraries, tutorials, workshops and blog posts – will help others understand when and how to apply these tools to their collections.

The potential of this project goes beyond creating metadata about digitised items: it’s a chance to apply data science methods to look for patterns and gaps that are visible only when you work across millions of pages with a broad temporal and geographic scope, while collaborating with historians to make sure the results can inform their historiographical work. 

Commercial tools aren’t always suitable for historical, scientific or art collections, and academic tools might be significant in one field but produce results that seem obvious to another. This project is a unique opportunity to work across disciplines to create results that could only come from radical collaboration.
 
Mia Ridge is the digital curator at the British Library

Comments