Tooling Up for Digital Humanities

A visualization can take any number of forms. While some will be geographic and map-based, others will involve various ways of displaying quantitative data, such as a simple graph, a word cloud, or a visualized network.

The Spatial History Lab at Stanford has experimented with a number of visualization technologies. While most are geography-based representations, others, like this display of human social networks (rendered in Pajek and published with Adobe Flash) show the value of other visualization tools.

Although many visualization technologies are often used merely as illustrations, the real power of visualization comes in its ability to make powerful arguments, and show data in a way that raises new questions. We may tend to think of visualization as a finished product, as part of presentation, but it may be more useful to think of visualization as part of the research process. Visualizing data, argues Richard White in “What is Spatial History,” is “a means of doing research; it generates questions that might otherwise go unasked, it reveals historical relations that might otherwise go unnoticed, and it undermines, or substantiates, stories upon which we build our own versions of the past.”

But conveying essential data in an elegant and clear way is harder than one might expect. There are plenty of visualizations that are appealing in some aesthetic way, but which might contain distractions and fail to convey the intended message. At worst, a poorly executed visualization can distort data and leave a reader confused.

The pursuit of clarity in visual representations of data has been one of the pursuits of renowned visualization scholar Edward Tufte. Tufte famously railed against the widespread use of “chartjunk,” which he argued was a distraction in conveying information as efficiently as possible. Others have defended chart junk as a way of engaging readers. An effective visualization requires a careful balance of design and function.

Visualizations offer new challenges in communication for another simple reason: it seems people are more willing to accept certain visual representations as objective or scientific. In a field in which scholars work with great amounts of uncertainty and ambiguity, what is the humanities scholar’s obligation in visualizing a data set? How must we account for visualizing data sets that may be incomplete or skewed?

The Republic of Letters project at Stanford, which is visualizing epistolary connections in the eighteenth century, has had to grapple with this problems. See linked video and interactive visualization.

In fact, visualizations are themselves fundamentally human creations, each one representing many small choices made by an author to show certain things and hide others; to emphasize certain features and not others. In that sense, visual arguments are like the forms of communication that humanities scholars are more accustomed to using, but it may be able to convey new information and raise new questions in a number of scholarly fields.