Effectively Visualizing Library Data

Abstract

Data visualization has become a hot topic over the last few years. This popularity can be seen everywhere. The New York Times revolutionized itself by creating gorgeous, interactive visualizations of everything from political campaigns to nutrition. (1) There is a suddenly vibrant data visualization blog scene, lead by stellar resources FlowingData and Information is Beautiful. (2) More and more software programs have appeared that make creating compelling visualizations easier and easier, but also more and more data are collected and analyzed digitally. Our new means of collecting data necessitates new means of representation to communicate the message of massive and messy sets of data. The immense success of computing in the second half of the twentieth century has brought about a correspondingly immense challenge: how do we deal with all of this data? Design thinking stepped up to the challenge by allying with statistics to create the half-art, half-science of data visualization. Libraries have not been left behind; everywhere the rise of new forms of data presentation is visible. The Seattle Public Library features a live-updating dashboard showing recent circulations and intricate networks of keywords. (3) The Harvard Library Lab is working on a powerful tool to view collection size and circulation by subject heading. (4) The Indianapolis Museum of Art has a web dashboard that boldly presents their changing quantities of artwork, memberships, visitors, and of course Facebook fans. (5) Brown University Libraries are working on a similar idea that presents live data such as checkouts in embeddable widget form. (6) The North Carolina State University Libraries have an ambitious data visualization project that aims to visualize the usage of reference services, course tools, computer workstations, and group study rooms. (7) These are all great strides being made by innovative libraries. Hopefully, as more examples appear and these frontrunners release their techniques as open-source code, the barrier to entry will diminish for all libraries. In the years to come, public data dashboards on library websites may be as common as catalog search boxes. It is not the aim of this column, however, to enumerate the most beautiful graphs of library data on the web but to demystify the practice of data visualization so that we can begin to create our own. THE PURPOSE OF VISUALIZATION Data visualization has a clear purpose: to aid in our understanding of data. Visualizations help us recognize otherwise obscure trends. Many visualizations offer means of interaction, they do not merely report a single, unassailable truth, but give the end-user an opportunity to explore the data and reach possibly unanticipated conclusions. They help us simplify the interpretation of an intimidating universe of data, intimidating both in terms of the sheer size of datasets, which can consist of millions of data points but also in terms of the dimensions of the data. A dataset might have only a dozen distinct points, but each point could in turn have thousands of aspects to it, a veritable world unto itself. Think of library branches: even large public library systems may have a single digit number of branches, but each of those branches has distinct sets of opening hours, geographic coordinates, items in the collection, staff members, and services. Distilling those differences into a meaningful representation is a challenge. [FIGURE 1 OMITTED] Data visualization approaches the challenges of size and complexity by fusing the art of design with the logic of statistics. Success hinges on both; a brilliant statistician can produce incoherent spreadsheets, while a talented designer can create misleading visuals. Above all, visualization strives to accurately represent data; mere artistry descends into chart junk and delusion. Figure 1 is representative of everything data visualization seeks to overcome. …