GIS and the Humanities, part 2

In the first part of this post, I described how a lot of ways in which work in the humanities is interacting with the spatial is in the process of generating “flat maps.” That is, they reproduce what is already in the texts themselves, without pushing any analytical balls forward.

These sorts of projects engage in the creation of what Deleuze and Guattari call “tracing.” Against the tracing, they put the “map,” which is a rhizome:

What distinguishes the map from the tracing is that it is entirely oriented toward an experimentation in contact with the real. The map does not reproduce an unconscious closed in upon itself; it constructs the unconscious. It fosters connections between fields, the removal of blockages on bodies without organs, the maximum opening of bodies without organs onto a plane of consistency… The map is open and connectable in all of its dimensions; it is detachable, reversible, susceptible to constant modification… [It has] multiple entryways and exits and its own lines of flight.

I imagine that the creators of these projects don’t see their work in this sort of negative way. Or, more importantly, they don’t care: they are creating intentionally pedagogical, not analytical, tools. That possibility makes me temper my own argument a bit, since I worry that I’m beating up on strawmen. For example, I still don’t know how to read the “too soon, far too soon” in considering GIS in one response to the Chronicle‘s interest in “Literary Geospaces.” But one way or another, These sorts of projects are the very tippity tip of the possibilities available in pushing forward an actual spatial turn away from the hegemony of history in the humanities. They acknowledge that space is important. The next step is unfreezing space.

What follows is an “annotated version” of the second part of a presentation I made this week on using GIS in the humanities (pptx | video), where I actually start talking about a GIS. All links open in new windows.

Simply put, a GIS is a system (the “S”) of information (the “I”) with a geographical (the “G”) component. We interact with a GIS whenever we ask Google Maps for driving directions: Maps interacts with a network GIS that figures out how to get from one point to another and then displays that information in our browser.

The most important conceptual idea to grasp of how a GIS is different from the various projects in the earlier part of the presentation is that we can ask a GIS a question and then get an answer. This could be driving directions, or it can get more detailed, like asking, “how are median ages distributed in the 60637 ZIP code, according to the 2000 Census?” or “what kinds of governments were in Latin America over the course of the 20th Century?” or “what is the relationship between the railroad and population distribution in industrial England?”

My own question in my workshop presentation was, “how can a study of the distribution of the various points mentioned by John Dos Passos in certain sections of the U.S.A. trilogy change our understanding of his representation of the US as ‘the speech of the people‘?”

The point here is that using a GIS can open analytic entryways into understanding an object of study. A GIS begins the engine of analysis, not stops it, as a flat map does, assertions about putting the reader into a text’s “imaginative landscape” notwithstanding.

U.S.A. trilogy vs. 1920 Census data. (click to enlarge)

Creating a GIS, though, is a rather tricky–or at least time-consuming–procedure. The first part is collecting data. For my project, I had data come from two sources. The first was the U.S.A. trilogy itself, where I hand-entered geographical observations and then geocoded them (using Google Earth as my gazetteer). I matched the geographical observations against US Census data from 1920, which I downloaded from the NHGIS site. NHGIS not only offers giant data tables (spreadsheets) with historical census data, it also provides historically accurate shapefiles for each census, which allow me to represent the census data spatially along with my observations about Dos Passos. One result of the comparison is something like this map here, which contrasts the distribution of sections of the trilogy with the US population as a whole.

What kind of conclusion is immediately apparent? Dos Passos created a much more widely dispersed US than the population data from the time would have dictated. This suggests that “speech of the people” wasn’t enough of the US for him–the geographic expanse of the US was also important in how he rebuilt a US in the trilogy, which then moves the ball further toward a larger, material claim I’m making about national imagining in the novel as a genre. This kind of conclusion would be rather impossible without the use of a GIS.

Once the hurdle of data entry is met, however, the next step is using software to interact with the data. In my talk, I mentioned GeoDa, which is a free, Windows-only geostatistical application. It’s very basic in its interface, but it can do quite a bit in comparing various variables and doing regressions on the variables. A neat little program. I also mentioned Quantum GIS, a far less exciting, but still free, piece of software. It’s less exciting mostly since qgis can’t do the sorts of statistical analyses that I would like my software to be able to do. Google Earth, of course, can also be used as something like a GIS, as shown in this movie by Matthew L. Jockers that shows the distribution of Irish-American novels’ plots (breathlessly reported by the Chronicle), but it does not have a lot of the analytical tools built in that GeoDa, qgis, or other programs do.

Once money gets on the table, ArcGIS becomes the sort of gold-standard for building a GIS, despite its amazingly frustrating interface. ArcGIS is useful to learn since it can do pretty much everything, and it is a good introduction to concepts surrounding GIS. I used ArcGIS to create the map above, for example. But it’s still annoying, expensive, and often completely illogical. At least the results it generates are decent.

At the University of Chicago, there is a small GIS community growing around a few classes about using GISes and a site license for the ArcGIS software. I myself am a part/product of that community, having taken the yearlong GIS course sequence offered by the Committee on Geographical Sciences. But beyond the relationship with the Oriental Institute, this sort of spatial connection has not permeated the Division of the Humanities. That’s too bad, but it might change.

Moving from my own academic setting back into using GISes in the humanities as a whole, it becomes clear quickly that there’s not terribly much out there, especially in fields connected with literary analysis. There are tentative first steps in the work of Moretti or Jockers, but beyond that it gets somehow far too connected to historical analysis–humanities GIS often piggybacks on historical GIS, even in my own work.

But if we keep Morettis words in mind:

A good map is worth a thousand words, cartographers say, and they are right: because it produces a thousand words: it raises doubts, ideas. It poses new questions and forces you to look for new answers.

We can crack open two new dimensions of analytical approach. That sounds exciting, since it is!

There were a few additional spatial links for the Humanities that I did not include in my presentation, but that I think could give possible avenues for future work.

• GIS & the Humanities from Stanford’s library page. I’m not surprised that two of the academics I’ve mentioned in this post are at Stanford and that they also have a straight up link for GIS and the humanities. Included is a neat little pdf that helps make the same kinds of claims I’m trying to make here.
• Mapping the Digital Humanities from HASTAC. This is a discussion that just began on mapping and the humanities. I don’t want to say too much about it.
• GIS for the Humanities from San Antonio College. I haven’t fully investigated this site.
• CrimeStat from the University of Michigan. Despite the name, CrimeStat can be used on any sort of point-based data (like events in a novel) to show deeper clustering and other relationships beyond what ArcGIS can do. It is free and Windows-only.

Tags: Deleuze and Guattari, digital humanities, Franco Moretti, geography, GIS, humanities, John Dos Passos, Stanford, tracing