This June I was grateful to receive a scholarship to attend the Digital Humanities Summer Institute, held at the University of Victoria in Canada. I first attended in 2017, whilst I was still studying for my Master’s degree in Book History and during the very early stages of researching asylum libraries. This time round, I had more of a handle on my project and its needs, so it was great to go back with a real purpose.
DHSI is a wide community - it brings together historians, librarians, literary scholars and academics from other disciplines - and it’s sometimes affectionately referred to as ‘summer camp for nerds’. It’s a great place to learn about the potential that Digital Humanities holds for research, and it also forms a hub for some very important discussions about approaching these methods in ways which are ethical and accessible, and acknowledge the many inequalities which currently exist. (The #RaceDH, #FemDH, and #QueerDH class hashtags on Twitter form an excellent resource).
I started off with Making Choices About Your Data, an excellent introductory class from Paige Morgan and Yvonne Lam. I spent most of my week working on data for the Asylums Map, tidying up and exploring some mapping software which might make it a little more stylish than good old Google Maps might be able to (watch this space). My second week was spent learning about Network Analysis with Jessica Otis, who introduced us to (some of) the mathematics behind networks, and how they might be useful for humanists. We got through the maths, we figured out how to structure data properly for network analysis, and we practiced collecting data and making networks using Samuel Pepys’ diary as a source. We chose a single week to work with and mapped Pepys’ social network. From Monday 23rd June to Monday 30th June 1662, Pepys did manage to get some work done - but he also spent a fair bit of time gossiping over drinks, getting boats around London, and having musical meet ups with his friends. He also encounters an unpleasantly wormy fish at dinner, and in true Pepys style, sexually harasses an employee. My classmate, Julia King, posted the network we made in class on Twitter (see left!) When I got back to my dorm room that evening, I couldn’t help but continue messing around with the data - looking at directionality and coding the nodes with Pepys’ social relation to the person. Gephi wasn’t having it that day and the colours are a little incorrect, but overall a really fun way of visualising social networks in a way that’s difficult to do in plain writing.
Each edge’s colour represents a different kind of social interaction; each node’s colour represents a different social relationship Pepys holds to that person. Edge weight (and arrow size) represents the number of interactions between Pepys and that person; edge arrows also represent the directionality of the interaction (for example, whether someone is gossiped about or does the gossiping, owes or is owed money.)
As anyone who works on asylum history knows, there’s a lot of potential for collecting data. The nineteenth century loved statistics (though often in an extremely problematic way). Asylum doctors collected a wealth of information about their patients, filtered through their own preconceptions, and took great care in collating and presenting it. This can be very useful for those of us who want to study asylums, but sometimes those numbers need to be taken with a pinch of salt. However, it’s not just the numerical data which is interesting for Digital Humanities techniques. Social networks, like Pepys’ above, show that qualitative data can work well too.
As I’d worked on the Asylums Map data during the previous week, I’d been reminded of some patterns which I thought might be nice to produce networks of: one of these patterns was the connections between asylums and their architects. I took a small sample of English asylums to work with, and produced a small network showing the overlap between architects and asylums. It’s a pretty rough visualisation, but it does pick out some of the major players in nineteenth-century asylum architecture - for example, Albert Edward Gough and John Giles, who worked together, and George Thomas Hine. Others, such as William Lambie Moffat, would pop up as more data is added to the network - Moffat designed the Stafforshire County Asylum in England, but later designed several Scottish asylums including the asylum at Montrose.
Pink nodes represent architects, green nodes represent asylums. Here, node size represents the ‘degree’ - i.e. how many connections the node has to others. I’d have liked to only apply this to the architects’ nodes.
My next plan is working on visualising the network of asylum libraries. Which periodicals, books and authors pop up most frequently in asylum libraries? As I’m working with my data, I can pick out a few that I think will be influential (Scottish asylums are stuffed full of Walter Scott) - but it’ll be interesting to run some analysis on the data too.
Many thanks to the DHSI organisers, our generous instructors, and the community who made it a productive couple of weeks!