Looking at a program from the Digital Humanities 2012 conference, one trend in digital humanities research that emerges is that of creating GIS layers from geographical information in historical texts. In the last few years, GIS has proven to be an incredibly useful platform for spatial analysis. However, one limitation is that GIS-compatible file formats have only been produced since the early 1990s. In order to look at data over a longer time period, researchers must digitize geographic information contained in print sources.
One work in this area from the DH2012 conference is “Content Creation by Domain Experts in a Semantic GIS System” by Alexander Nakhimovsky and Tom Myers. This idea of this work is to create a repository of historical event maps (such as country boundaries) that would host the results of digitization efforts performed by other researchers. The database supports many different formats, from shapefiles to KML, which allows content at different levels of detail to be included. The main goals seem to be visualization and database-building. Visualization is a good use of historical data because some of the problems (level of accuracy, compatibility with digital-age datasets) with digitization are less important.
Another presentation from the conference, “Uncertain Date, Uncertain Place: Interpreting the History of Jewish Communities in the Byzantine Empire using GIS”, deals with the issue of how to represent uncertainty of data. The author, Gethin Rees, suggests that maps have a higher perceived authority than text, and that scholars are less rigorous in checking source data when referencing or reprinting a map. In a GIS layer, data points must have an exact geographical location, but in some cases the historical documents give a general area as opposed to an exact location. Rees has decided to represent the uncertainty of the location through the transparency of the map symbol. A more transparent symbol indicates less confidence in the location.
A third paper from the same conference, “Ptolemy’s Geography and the Birth of GIS”, analyzes the data structure of Ptolemy’s Geographike hyphegesis. Ptolemy’s work mapped the entire known world and provided tables of cities with geographic coordinates. Leif Isaksen created a histogram of the minutes in latitude and longitude for all the cities included. Isaksen shows that 0 and 30 minutes are by far the most common (representing whole and half degrees, respectively), suggesting that at least some of the coordinates were approximate while others are clearly more precise. Furthermore, his analysis shows that nearly all of the high-precision locations (coordinates not ending in 0 or 30) are in Hispania, Italia, Greece, Asia Minor, and Egypt while the more distant regions of the known world had no high-precision locations. Isaksen presents this finding, as well as graphs showing the order in which interior cities were listed, as proof that Ptolemy compiled the city locations from different source data. This result demonstrates that the precision of data can be variable, even within the same historical document. His findings could be used to calculate statistical measures for matching the coordinates to modern archaeological sites or representing the city locations alongside modern data in a GIS setting.
These three papers approach the same topic, pulling geographical information from historical works and putting it into GIS layers. Since they are looking at historical data for different reasons, so the methods are not really comparable. Nakhimovsky and Myers are creating a database of visuals to accompany the historical narrative, while Rees and Isaksen are working on qualitative and quantitative representations of data precision, respectively. However, combinations of these approaches are conceivable. Some representation of data quality, whether it resembles Rees’ approach, Isaksen’s approach, or something in between, would clearly be valuable as metadata for databases like the one created by Nakhimovsky and Myers.