My research focusses on medieval place-names and what they tell us about languages, landscapes and cultures in the past. I attended the workshop as I wanted to explore the potential for using machine learning to identify topographical features from maps that have been digitised. Being able to identify landscape features more quickly than is possible if manually checking maps would allow us to ask different questions about naming. Instead of asking ‘how is this word used in naming?’, we would be able to ask ‘how can feature X be named?’ I wanted to find out whether machine learning tools would make identifying features possible at scale, to enable this second analytical approach to be used. Having attended the workshop, I think the second approach might be possible be and am intending to carry out a pilot study of naming of wetland areas using this approach. (Once I’ve checked that a similar dataset doesn’t already exist...)
For me, the workshop served as an introduction to Python and machine learning as well as the tool we were using, MapReader. The workshop was designed so as to be appropriate to people with a variety of existing background knowledge. As someone with no Python experience, the provision of code in Jupyter Notebooks – with some sections to be completed or edited by attendees – enabled me to participate, and see how I could edit the code to adapt it to my own research. Some participants with more coding experience applied the tool to material of interest to them throughout the workshop, and I was able to do this at the end of the workshop. I also feel much better informed about how machine learning is carried out than I did before attending the workshop.
University web-page: Eleanor Rye