Library Observatory: an open, community facility for exploring library collections and services

In connection with our emerging Data Artifacts initiative to understand the impacts of data in networked experience, we’re eager to explore some of the most culturally-determined troves of information our society produces: the bibliographic datasets that comprise library online public catalogues. We realize, of course, that one of the richest troves of such information lives in our institutional backyard, associated with the collections of the Harvard Library. Despite extensive consolidation, there are over seventy library units at Harvard, each with its own extensive collection. HOLLIS, the Harvard Online Library Information System, allows patrons to search for select volumes, build lists, and export them to other media, but it does not afford panoramic views of the entire holdings that reveal macroscopic patterns in the acquisition, distribution, circulation, and citation of the university’s collections over time.

To explore these data—and ultimately, similar datasets with other institutions—we’re starting the Library Observatory to create and host a shared set of visualization tools for contemplating the full range of open library information. It’s an engagement made possible thanks to the work of our colleagues in the Library Innovation Lab to make Harvard Library data open and programmatically accessible through systems like Library Cloud and the open API project of the Digital Public Library (DPLA). We are seeking initial funding to develop an open-source library graphics code base—the technical framework for the Observatory—and a sample set of highly usable, intuitive visualization tools tailored for library collections in all their technical, socio-cultural, and material dimensions. Crucially, we plan to express and develop these tools in an open forum for discussing, contextualizing, and refining our approach, in conversation with others involved in collections-driven scholarship, teaching, and administration.

Thus, the Observatory will be technologically-enabled, but also sensitive to the ways in which information is always embedded in social, cultural, and institutional contexts. Information wants to be free, but it needs to be in dialogue—with the people that harvest it, the institutions and social relations that structure it, and the audiences that activate it. An astronomical observatory is more than a telescope; it encompasses the entire site where the instrument is used and observations are made, shared, and debated. In similar fashion, the Library Observatory will develop not only tools and instruments, but discourses: first through one-on-one interviews with librarians, then in group workshops with the extended Harvard community, and ultimately through external, online networks of scholars and students. The Observatory has broad implications, not only as a tool to explore Harvard’s collection practices, but as a model and source of support for the global communities tracing the varied trajectories of open library data.

We have already made initial headway in the construction of the library graphics code base—a set of our own open-source algorithms for translating library records into graphic traces on screens and on the web. The code base is being developed by Travis Bost (MDes ’12)—architect, terrifically talented data-visualization designer, and alumnus of metaLAB’s curatorial innovation fellows program—who is also devising a growing set of visualization tools using Processing. These interactive representations present cross sections through the Harvard collections, prompting questions and discussion about what open library data can tell us about the changing nature of scholarship at the university, its collecting habits, shifts within and across media types, and the like.

The following images offer examples of the look and feel of these custom-designed data-visualization tools and the promise they offer for sparking research questions, refining library practice, and producing teachable moments.

Here’s a still frame from an animation charting the evolution of early-modern printed works at Harvard by their place of publication and by total number of volumes (click here for the full animation):

Two comparative timelines showing accessions of non-English items at Harvard over time from 2002 to 2012 on the lower line correlated with dates of publication, from 1750 to 2000, above:

A map of all public libraries under the Institute for Museum and Library Services showing relative intensity of borrowing activity normalized by total number of registered borrowers:

In one-on-one interviews and group workshops, gestural data sketches like those above will be tweaked, refined, annotated, and contextualized. Such framing dialogues and conclusions drawn from them will be captured for online publication; they will also help to build better tools and analyses, and to further the project of bringing open data into dialogue with research, teaching, and library work.

The Library Observatory project emerged from a series of discussions with library staff in the context of metaLAB’s longstanding interest in the technology, design, and cultural history of libraries—the role the library plays both as an institution and as a metaphor for knowledge, learning, and memory. As we build a visual language and interface for seeing Harvard’s collections from a variety of perspectives, those conversations are evolving. Exchanges around rich visual representations of open library data elucidate new questions about the history and future of Harvard’s collections and challenge our assumptions about the authority of data in the lived library world.

We have a chance to explore that kind of exchange tonight, at metaLAB’s openLAB event: Travis will be demonstrating a number of his developing data-visualization tools, and showing off a wonderful interactive bookshelf he’s cooked up as well. Join us if you can, and help us think through ways to open the Observatory to the diverse constituencies of the Digital Public Library of America and beyond—to encourage libraries to open their own data to analysis; to foster a nuanced appreciation for the historical and social embeddeness of libraries in transition; to prepare a moveable feast of data in dialogue.