How Libraries can support Digital Humanities: reflections on #GaleDHDay

By Antony Groves (Learning and Teaching Librarian at the University of Sussex) @AntonyGroves

At the beginning of May, Gale organised their first Digital Humanities Day at the British Library. The event brought together a diverse range of speakers from around the world who spoke about different aspects of Digital Humanities (DH) scholarship; from infrastructure through to research and teaching. This post will draw out three themes from the day in an effort to better understand how we can support this growing area of work:

1. Collaboration in DH research is key and libraries can play a role within these collaborations. 
2. There are many different datasets, techniques and tools being used yet a common approach we can take to developing training.
3. We should work on our own data projects if we wish to really understand what is needed to support the academic community.

Collaboration in DH research is key and libraries can play a role within these collaborations.

In the afternoon session, Dr Sarah Ketchley stressed that “Digital Humanities projects are inherently a collaborative undertaking” and the earlier presentations of Professors Mark Algee-Hewitt and Joris Van Eijnatten highlighted this. The work done by Prof Algee-Hewitt and others at the Stanford Literary Lab has involved a number of ‘distant reading’ projects where participants have used a variety of computational techniques to analyse large collections of digital texts. Looking at grammar and language respectively, Prof Algee-Hewitt’s research involved digital novels whereas Prof Van Eijnatten focused on newspapers using The Times Digital Archive; both resources that libraries can provide.

Throughout the day, flags such as these indicated potential roles for libraries in DH collaborations. For example, Dr Julianne Nyhan reflected on infrastructure and the challenges to researchers of obtaining data in a format that can be ‘mined’ – in one case having to obtain a hard drive from a provider. This is somewhere librarians can help and Lisa Mcintosh, Director of Access Services at the University of Sydney Library, shared an impressive list of services offered by their library in support of digital research:

• Provide content for text and data mining
• License permission and copyright support
• Recommending tools and TDM (Text and Data Mining) resources
• Integrating text mining into Information Literacy classes in the Humanities
• Assisting humanities teaching staff to integrate text mining in the classroom
• Getting started with data visualisation training • Data analysis and visualisation guide

There are many different datasets, techniques and tools being used yet a common approach we can take to developing training.

For those wondering which students this area of scholarship might appeal to; the answer is all of them. In an inspiring talk about introducing DH in the Undergraduate Classroom, Dr Sarah Ketchley showed that her 2018 ‘Introduction to Digital Humanities’ module was full, with 35 students from 21 different departments across campus. Not only is this type of scholarship appealing to students but it is also invaluable to them. For one reason, as explained by Dr Melodee Beals, “evidence is merely data with a direction”. If we want students to critically engage with evidence-based research, helping them to analyse the underlying data is of great importance.

The tools that students use in Dr Ketchley’s class have included OpenRefine, Voyant Tools and more recently the Gale Digital Scholar Lab – a cloud based platform containing a range of software that can be used with Gale databases to which the institution subscribes. This cloud based approach avoided issues encountered by previous cohorts where a whole lesson had to be dedicated to downloading and installing the required programs. Dr. Tomoji Tabata also introduced an open source tool called Stylo to be used for ‘rolling stylometry’, a technique to detect stylistic changes in passages of text.

Throughout the day, reference was made to many different techniques (e.g. topic modelling, named entity recognition, sentiment analysis); tools (e.g. Gephi, Google Fusion Tables, MALLET); and data sources (e.g. TROVE, Hathitrust, Gale Historical Newspapers). With so much out there, it can be hard to know how best to start providing support. Thankfully, Associate Professor Ryan Cordell brought clarity to this undertaking by proposing four steps to teaching humanities data analysis:

• Start with creativity 
• Teach using domain specific data 
• Foreground corpus over method
• Foreground mind-set over method (‘programmatic’ thinking more important that programming’)

We take a similar approach to developing our Information Literacy training sessions and find that it works well. In the short amount of time that we often get to see students in workshop, making the content of the session as relevant to a given cohort as possible increases engagement. In addition, focusing on how to approach searching (as opposed to how to use a particular tool) means that they can apply this learning to a range of tools that they may encounter not just the one or two included in the session.

“Work on your own data projects to understand what is really needed to support your academic community”.

This is a direct quote from the final presentation by Lisa Mcintosh, which was the perfect way to finish the day. While listening to the research presented throughout the day was fascinating and certainly highlighted areas where we can support this scholarship, managing our own data projects and facing the same barriers that our researchers encounter is what will really help us to understand the support that is most needed.

This may sound daunting but hopefully this post has shared at least a few resources that can be explored further, and take encouragement from Prof Van Eijnatten who asserted that “if I can write a few lines of code anyone can”.