30 June 2023
Christina Nguyen
Graduate Student,
University of Toronto
UKSG 2023 was, for me, 2.5 days of non-stop intensive learning, with an excellent pinch of socialization thrown in for good measure. As a recently-graduated student from the M.I. degree, I found that attending UKSG provided me a practical understanding of the ecosystem of actors in the scholarly communications community, and pushed me to learn the vocabulary of the field very quickly. One of my main gripes with the degree is the overemphasis on theoretical explanation of concepts (for example, open access) but no teaching of how a librarian uses it day-to-day (e.g. signing transformative agreements with publishers). To me, both visiting the exhibitor booths and hearing the speaking sessions were incredibly valuable parts of the conference.
There were three breakout rooms that I particularly enjoyed; the first was Certain in our certainty (about working with limited resources as collections development librarians, particularly when shifting to the new landscape of transformative agreements); the second was A text data mining journey from the University of Birmingham (about plans and challenges to scaling up their text data mining supports in the library, from a practical perspective); the last was It’s in the data: improving student outcomes with digital content (a case study at Coventry University using the BibliU e-textbook platform to collect data and help students succeed academically). Each session offered something for me in the short- and long-term view of my career; then the plenary sessions helped me develop my values of the profession even further.
Certain in our uncertainty was particularly valuable for me to understand the day-to-day concerns of collections librarians from two perspectives: from a smaller university with limited funded, and a larger one with high amounts of collections to purchase. The starting assumption was that we live in an ecosystem that is rapidly transforming to read-and-publish (transformative agreements) models, so how do we (as librarians) see the challenges of this shift? As I had just graduated from the M.I., and not yet properly worked as a full-time librarian, this interactive session was an excellent insight into how other librarians’ day-to-day problems arose and were dealt with, including things like granular levels of budgeting and calculating the amount of savings for the university if paying for specific journals vs. buying a third-party-curated bundle. Because I have never been involved in collections development in my two years of the M.I., nor been able to take a practical course about this, Certain in our uncertainty was almost like a crash course on how different universities can approach the shift to TA. Hearing from both Liverpool John Moores and Manchester in the same session was wonderful as they faced some very different challenges to each other, but also then learned from each others’ experiences to improve their own challenges. I learned many tips and tricks (like Manchester’s model for making decisions on when to buy a transformative agreement) that I hope to use if ever I get the chance to work as a collections development librarian. I highly suspect, and hope, that future years of UKSG conference will have more case studies like this, where two or more universities collaborate in the talk to learn from each others’ lessons in collections development re: TAs.
A text and data mining journey was immediately practical and useful for my development as an early career librarian. As a student librarian in the digital scholarship department at the University of Toronto Mississauga campus, I am very keen on understanding how we can upscale our own supports for text mining. Our other campus’ libraries – we have 40 in total over three campuses – have all their own varying degrees of support for text mining, almost working semi-independently on their own powers. Though as a student librarian, I have little power to change the infrastructure (rather, I am part of the infrastructure as dictated by others), I can step back and see how that infrastructure can be changed or added to so that we can radically increase the amount of research our campus does with text analysis. The major premise that Birmingham acknowledged was that all text and data mining research at a university begins with library support. This is because the TDM method is a method not a field of study, and therefore there is not always a liaison librarian or some area expert for researchers to turn to for practical research skills help. Our role as librarians is to provide all of that infrastructure, which I expanded in my mind to be critically including:
- staff who can teach multiple levels of workshops in text mining, in different computational languages and with different fields’ approaches, such as
- in Python, R, and MATLAB
- for undergraduates, graduate students, and early-career researchers
- from a social sciences data science perspective, from a digital humanities data science perspective, and so on.
- a place to store, both physically and digitally, the corpuses that the library has purchased or gained access to for the researchers. For example, if a researcher has requested that the library purchase the data from X publisher, and the publisher has provided it in the format of a hard drive, what happens to the hard drive during and after the researcher has used it? How do you manage the multiple copies of the data floating around the university ecosystem? What are the legal challenges of accessing certain data types?
I suspect all of the questions Birmingham raised were immediately relevant to the current state of text mining work and infrastructure and UTM library. As a smaller library compared to our other branches on the main campus (UTSG), there is much scaling-up to do; and as a student librarian I hope to enact some of these changes in the short-to-long term. Canadian standards already exist for some of these challenges (for example, we have the Borealis platform, based on the American Dataverse platform, for storing research data with many different options for sharing and preservation) so it is my job now to look into how I can integrate the solutions Birmingham proposed with Canadian standards; if there are no Canadian ways, then optimistically, perhaps I ought to begin making a way with advice from the lessons I learned at this session.
It’s in the data is one of those talks that I will be processing for a while before I fully understand its value. Though it appeared to be a simple case study of an early-successfully implemented ebook textbook system (to improve students’ learning), there were many layers of discussions hidden in the main case, which the speakers acknowledged but didn’t quite get enough time to discuss. For example, from my extrapolations, if this model (of an ebook aggregator, like BibliU) was implemented at another university library, questions of this kind need to be asked:
- How do we make good sense of the statistics that we collect? How can we recognize the limitations and power of those statistics? Speaker Phil acknowledged that he was not a statistician, and that there were some potentially faulty reasonings behind the interpretation of the numbers.
- How can we evaluate what our faculty and students need in a platform like BibliU? How do we keep maintaining and updating those platforms to be flexible to those needs? What information practices and behaviors are occurring in our classrooms and outside re:textbook usage?
- What ethical challenges are there in collecting, anonymizing, and using that data in a sensible manner?
- How can we use this data in combination with other known facts about student needs to develop a larger, context-driven understanding of where library services need to go?
All in all, UKSG 2023 was chock-full of lessons that I am still, as I write, thinking through and finding more layers in. The experience has been very formative to my thoughts and understanding of the librarianship profession, and I am grateful to have been able to attend.