Libraries & Research: Supporting change in research

[This is the second in a short series on our 2014 OCLC Research Library Partnership meeting, Libraries and Research: Supporting Change/Changing Support. You can read the first post and also refer to the event webpage contains links to slides, videos, photos, Storify summaries.]

Anja Smit (University Librarian at Utrecht University) [link to video] chaired this session which focused on the ways in which libraries are or could be supporting eScholarship. In opening she shared a story that reflects how the library is really a creature of the larger institution. At Utrect the library engaged in scenario planning* and identified their future as being all about open access and online access to sources. When they brought faculty in to comment on their plans, they were told that they were “going too fast” and that they needed to slow down. Sometimes researchers request services and sometimes the library just acts to fill a void.  But innovation is not only starting but also stopping. The Utretch experience with VREs are an example of a well-reasoned library “push” of services – thought they would have 200 research groups actively using the VRE but only 25 took it up. Annotated books on the other hand is an example of “pull,” something requested by researchers. Dataverse (a network for storing data) started as a service in the library that was needed by faculty but ultimately moved to DANS due to scale and infrastructure issues.  The decision to discontinue local search was a “pull” decision, based on evidence that researchers were not using it. Ultimately, librarians need to be “embedded” in researcher workflows. If we don’t know what they are doing, we won’t be able to help them.

Ricky Erway (Senior Program Officer, OCLC Research) [link to video] gave her own story of push and pull — OCLC Research was asked by the Research Information Management Interest Group to “do something about digital humanities”. The larger question was, where can libraries make a unique contribution?  Ricky and colleague Jennifer Schaffner immersed themselves in the researchers’ perspective regarding processes, issues, and needs, and then tried to see where the library might fill gaps. Their paper [Does Every Research Library Need a Digital Humanities Center?] was written for library directors not already engaged with digital humanities. The answer to the question posed in the title of the paper is, “It depends.”  The report suggests that a constellation of engagement possibilities should be considered based on local needs. Start with what you are already offering and ensure that researchers are aware of those services. Scholars enthusiasm for metadata was a surprising finding — humanities researchers use and value metadata sources such as VIAF. (Colleague Karen Smith-Yoshimura has previously blogged about contributions to VIAF from the Syriac scholarly community and contributions from the Perseus Catalog.) A challenge for libraries is figuring out, when to support, when to collaborate, and when to lead. There is no one size fits all in digital humanities and libraries — not only is it the case that “changes in research are not evenly distributed,” but also every library has its own set of strengths and services which may be good matches for local needs.

Adam Farquhar (Head of Digital Scholarship at the British Library) [link to video] talked about what happens when large digital collections are brought together with scholars. Adam’s role, in brief is to get the British Library’s digital collections into the hands of scholars so they can create knowledge. Adam and his team have been trying to find ways to take advantage of the digital qualities of digital collections — up to now, most libraries have treated digital collections the same as print collections apart from delivery. This is a mistake, because there are unique aspects to large-scale digital collections and we should be leveraging them. The British Library has a cross-disciplinary team which is much needed for tackling the challenges at hand. Rather than highlighting the broad range of projects being undertaken at the BL, Adam chose instead to focus on a few small, illustrative examples. In the British Library Labs, developers are invited to sit alongside scholars and co-evolve projects and solutions. The BL Labs Competition is a challenge to encourage people to put forward interesting projects and needs. Winners of the 2014 competition included one from Australia (showing that there is global interest in the BL’s collections). One winner is the Victorian Meme Machine, which will pair Victorian jokes with likely images to illustrate what makes Victorian jokes funny. Another project extracted images from digitized books and put a million images on Flickr (where people go to look for images, not for books). These images have received 160 million views in the last year. These are impressive metrics especially when you consider that previously no one alive had looked any of those images. Now lots of people have and they have been used in a variety of ways, from an art piece at Burning Man, to serious research, to commercial use. Adam’s advice? Relax and take a chance on release of information into the public domain.

Antal van den Bosch (Professor at the Radboud University Nijmegen) [link to video] spoke from his perspective as a researcher. Scientists have long had the ability to shift from first gear (working at the chalkboard) to 5^th or 6^th gear (doing work on the Large Hadron Collider). Humanists have recently discovered that there is a 3^rd or 4^th gear and want to go there. In the humanities there is fast and slow scholarship. In his own field, linguistics and computer science, there is no data like more data. Large, rich corpuses are highly valued (and more common over time). One example is Twitter – in the Netherlands, seven million Tweets a day are generated and collected by his institute. Against this corpus, researchers can study the use of language at different times of day and use location metadata to identify use of regional dialect. Another example is the HiTiME (Historical Timeline Mining and Extraction) project which uses linked data in historical sources to enable the study of social movements in Europe. Within texts, markup of persons, locations, and events allow visualizations including timelines and social networks. Analysis of newspaper archives revealed both labor strikes that happened and those that didn’t. However, library technology was not up to the task of keeping up with the data so that findings were not repeatable, underscoring the need for version control and adequate technological underpinnings. Many times in these projects the software goes along with the data, so storing both data and code is important.  Most researchers are not sure where to put their research data and may be using cloud storage like GitHub. Advice and guidance are all well and good but what researchers really need is storage, and easy to use services (“an upload button, basically”). In the Netherlands and in Europe, there are long tail storage solutions for e-research data. Many organizations and institutions say “here, let me help you with that.” Libraries seem well situated to help with metadata, but researchers want full text search against very big data sets like Twitter or Google Books. Libraries should be asking themselves if they can host something that big. If libraries can’t offer collections like these, at scale, researchers may not be interested.  On the other hand in the humanities which has a “long tail of small topics,” there are many single researchers doing small research projects and here the library may be well positioned to help.

If you are interested in more details you can watch the discussion session that followed:

I’ll be back later to summarize the last two segments of the meeting.

*A few years ago, Jim and I attended one of the ARL 2030 Scenarios workshops. Since that time, I’ve been quite interested in the use of scenario planning as an approach for organizations like libraries that hope to build for resilience.