Cartography: Historical GIS
This week’s main reading was Placing History: How Maps, Spatial Data, and GIS Are Changing Historical Scholarship, which came out of a 2004 conference at the Newberry Library. This volume should be recommended reading for any scholar of history, not just those interested in GIS, and should be used in historical methods classes.
Although one could take many angles in discussing the articles in this book, I’d like to focus on one of the methodological concerns that the book addresses: the disjuncture between the seeming certainty required for data entered into GIS software, and the uncertainty that is inherent in much of what we do as historians (as David J. Bodenhamer discusses in his essay “History and GIS: Implications for the Discipline”). GIS requires us to input data sets showing a degree of certainty, when, as those of us who took Clio 3 last fall can testify, our data sets (whether we or someone else created them) tend to be far from certain. Indeed, we spent much of Clio 3 not as much using data as getting it ready for different types of applications. Many of us struggled with how to display the uncertainty inherent in our data. For example, mySQL databases require an exact date (yyyy-mm-dd) for a value to be read as a date, instead of as a plain old number. Often we don’t have an exact date, just a year–if we’re that lucky. Entering just the year in a mySQL database, though, yields a date on the first of the year. Sometimes that’s fine, but often it isn’t. How do we convey that? As this indicates, uncertainty isn’t just the province of using data for GIS applications. In the end, the applications to which we could put our data sets outweighed the concern about ambiguity, something the authors of the essays in Placing History argue, but there were still those questions.
As of the publication of the book in 2008, those questions were still unresolved in GIS. I’m curious how much things have advanced in the time since. Are software packages like Neatline and d3 good ways to show ambiguity, or are they not viable alternatives to something like ArcGIS? What do others know about how the field has advanced in this time?
Addendum: This week I commented on Nate’s blog post.
This is a great post, David, and you bring up a really important point about the use of GIS for historical scholarship. Scientists have a phrase “garbage in, garbage out” meaning that you’re results are only as good as your data. Stick bad data in a great tool and you get a bad result. Of course, as you point historians deal with data where certain attributes, like the date created, are often ambiguous. We LIKE our garbage. We don’t even consider it garbage. So, then, how can we make use of tools that see our treasure as trash? I think you’re right that the answer lies in projects like Neatline where GIS has been developed specifically FOR digital humanists and not social scientists. Perhaps, historians working with GIS can also present their data in multiple ways — removing uncertain data points for one version of the GIS map and plugging them back in for another and then discussing the difference. It definitely shows that all technology has its own biases and perspectives — which, as you say, makes it a perfect topic for a digital methods class.