The two networking tools I used for this practicum — Palladio and RAW — were each unique in different ways. And, despite using the same Civil War data set for each one (battle and unit), I managed to glean different things from the same data by viewing it through the two different tools.
Palladio proved extremely easy to use. The drag-and-drop feature allowed me to load an Excel spreadsheet with the Battle and Unit data very quickly. The download box populated rapidly, and I was able to produce a graph very quickly with a simple “click.” I’m not sure that the floating nature of the graph was of much help to me, though. Granted, I could expand and contract its various nodes and edges quickly, but the graph was most useful to me in its static form.
Interestingly, the Palladio graph allowed me to see very quickly that the various units represented in the data, such as the 1st Michigan Cavalry and the 136th New York Infantry, generally fought the war in the same region. The 1st Michigan Cavalry stayed principally in northern Virginia as indicated by its visual linkage to battles such as Old Church, Winchester, Centreville, and Brentsville. By contrast, the 136th New York Infantry spent most of its time in the South, fighting at Atlanta, Chattanooga, and Stone Mountain. Yet the graph indicated that at some point, both units participated together in the battle of Gettysburg, suggesting that the “regionalization” of various Union regiments did not mean that the Army’s senior leaders could not call upon them to move and fight elsewhere. But the most significant thing I took away from the network visualization of these units and the battles in which they participated was that, for the most part, many of them fought in one general region within the United States and seldom moved from that area. Perhaps one explanation was the difficulty inherent in moving a foot-borne Army from one place to another quickly. Locomotives offered limited support, and damaged rail networks throughout the South likely complicated train traffic.
My only difficulty with Palladio was not with the program but with my ability to figure out how to import a screenshot of my graph into the body of my blog. I’m still figuring out how to do it. But, in the meantime, a pdf version of that screenshot appears at the following hyperlink: Palladio
Like Palladio, RAW was easy to use. The drag-and-drop upload feature resembled that of Palladio. The data uploaded quickly, and I was able to generate networking diagrams almost instantly using the Battle and Unit data set. I began with an Alluvial Diagram, which was very difficult to use, even after I adjusted the height and width repeatedly. The data lines were not easy to follow, but the tightly packed lines suggested, contrary to my Palladio graph, that many units fought in some of the same battles. For example, the varying thicknesses of the bars to the left of each battle name seemed to suggest a hierarchy of common battle participation among units. If I read it correctly, then this information was more useful than the Palladio graph, which did not really capture those commonalities in a clear, comprehensive manner. In addition, I was not sure what the various colors assigned to each unit were telling aside from possibly serving as a visual guide to lead me to certain units in the network graph much more efficiently. Here is the Alluvial Diagram I generated.
The next network I generated was a Circular Dendogram (whatever that is), which tended to reinforce the “regionalization” impression I got from my Palladio graph. In this case, though, the resulting graph was much, much clearer and easier to follow. Yet unlike Palladio, I was able to see more examples of units fighting in multiple regions. For example, the network diagram confirmed that the 4th New York Cavalry was strictly a regionally aligned regiment while the 136th New York Infantry, as suggested by the Palladio graph, fought in northern Georgia and later at Gettysburg. But the Dendogram helped me see that the 136th also fought cross-regionally throughout northern Virginia, most notably at places like Chancellorsville. This particular network diagram did the most to convince me that not all Union regiments were wedded to one region, further suggesting a higher degree of deployment capability and mobility than one might expect of a horse-drawn army. Here is how my Dendogram looked:
Overall, I found both tools somewhat useful in providing clues to the regional mobility of various Union regiments during the Civil War. But I’m not certain that these tools told me something that I couldn’t have gleaned from the Excel spreadsheet. Frankly, the hours of data construction that went into the spreadsheet was the real work. Uploading it and creating the network graphs was a breeze. The visualizations were certainly intriguing; but, at some point, my guess is that the data compiler could have come to the same conclusions about mobility and regionalization without graphically representing the data. Frankly, I prefer the the visualization and its impact, an impact made all the more effective by the general ease of use involved in reading the results of both programs.