Seeing the unseen

Seeing the unseen

Wednesday, June 5, 2013

Data visualization (data viz among the human-computer interaction community), which is making visible the relationships between sets of data, abounds in our hyper-kinectic and hyper-connected world. From the ubiquitous word clouds (tag clouds) to the all-too familiar maps of the United States political landscape (especially during elections!) to real-time traffic condition maps accessible on our smart phones, we might find it hard these days to get through a day without one good visualization. The really good data visualization draws the mind to meaningful relationships and conclusions one might otherwise never have discovered. A good data visualization also can take a known fact or concept and amplify its significance, often in haunting, beautiful and profound ways.

Data visualization has been an area of intense activity in universities for maybe two decades, but particularly the last decade there has been tremendous growth both in the scope and sophistication of projects of inquiry and in the reputation such centers of inquiry have earned, such as MITH here at UMD. You know you've arrived when there are bi-annual or yearly globe-hopping conferences to attend!

So where does Art History fit in all of this? How does a field that deals with the visual, with the created visible object, make use of data visualization? As you might imagine, the earliest forays in data visualization concentrated on bringing to the eye that which no longer is visible or visitable, often with stunning results. This data visualization is called modeling, but it is data visualization just the same, the best of the models made after painstaking research and study of relevant archaeological or archival materials. The Rome Reborn project, a joint IATH (UVa) and Visualization Portal (UCLA) venture, and Lisa Snyder's ongoing reconstruction of the The World's Columbian Exposition in 1893 in Chicago at the Urban Simulation Team (UCLA) are two excellent examples. Indeed, the Collaboratory's visualization space, with its large, curved screen, privileges the viewing of such immersive models. Beyond that, what is there in data visualization for Art History? Better asked, how might Art History make use of the powerful tool of data visualization?

Over the past three years here in the Collaboratory we have focused alot on creating maps and tours in Google Earth that could prove useful to teaching in the classroom (and look for some exciting, interactive developments with our extant maps sometime later this summer!). Matthew Lincoln, one the past Spring's graduate assistants in the digital humanities and working here in that capacity for the summer, saw intriguing possibilities in those early maps and their awkward interconnectivity, especially with specific placemarks, and set out to do something a little different, that is to visualize a primary text, Albrecht Durer's diary account of his travels to the Netherlands in 1520-21. An extremely detailed accounting of the artist's travels, the diary offers art historians a feast of data that when visualized can exist independent of the sequential narrative in which they are found and thereby take on new life and new meaning. One is not bound to the sequence of the diary itself but can explore via discrete nodes of time or places on a map, with the relevant pages from the diary or images of works mentioned readily available for quick consumption. This interaction is made possible by the fact that the primary text has been turned into a relational database by careful coding of the text so that, for instance, a placename is identified as something to be geolocated, and a work of art can be tagged and associated with a jpeg file of it. That relational database then can be translated into, in this case, a Google Earth visualization. More importantly, it can be translated in different directions, depending on the type of visualization one wishes. As well, because Durer's text now is a base table of cells of data, it can be merged with other databases, say of artists active around the time of Durer's trip, rendering visible possible connections not mentioned in the text.

As impressive as is this Spring project, Matt is not done. Currently he is trying to access XML tables of artists biographies (think Getty's ULAN) and integrate those into a data visualization of the movements of Dutch and Flemish still life artists over the course of their careers, an idea first conceived and tentatively mapped out in Google Earth by another graduate student in the Department, Kristi Jamrisko. If you happen to be at the THATcamp Prime  at George Mason University this Friday and Saturday, you'll undoubtedly get a sneak peek and can help Matt think through the challenges of visualizing data that is not always so precise (what does one do with circa dates, after all?). Matt's work is worth keeping an eye on, as is his excellent blog.