For my final project I worked on mapping Army posts in post-Civil War America. This project stemmed from my wider research interests, and I hope to use the insights gained in this project to further my dissertation research moving forward. For my dissertation I am looking at the professionalization of the Army in the last half of the nineteenth century. One of the factors often cited as part of this professionalization is the consolidation of units onto a smaller number of larger bases, allowing for both economies of scale in fatigue and guard details and the ability to train soldiers to operate as larger units. With this in mind, I wanted to create a map to help visualize the shift both in number of posts and the greater troop strength at each.
Like all projects, this one first had to start with sources. For this project to be effective I had to have a pretty complete list of Army posts as well as their assigned strengths and locations. Additionally, I would have to have this same information over the course of several decades in order to show change over time. While I did not have an existing source base at the outset of the project, I had a pretty good idea of where I could find the information from previous research. On an earlier research paper I worked with the “Annual Report of the Chief of Ordnance”, and I knew that this report, along with those from all of the other departments, was compiled annually and submitted as appendices to the “Annual Report of the Secretary of War to Congress”. A quick google search found the one year (1877) that had been digitalized by Google Books. Contained within the appendices was the report of the Adjutant General, which included a table listing all Army posts by name, location, commanding officer, units assigned, and strength (broken down in more detail than I needed for the parameters of this project). Unfortunately, only one year of the Report has been digitalized, so I had to make a trip to the Army Heritage Center at Carlisle Barracks to get the data for the other years. However, since I was already able to look at the 1877 report, I showed up knowing exactly what I was looking for and exactly where to find it, so it took me only a half day to pull the data from 30 years of Annual Reports.
Once I had the data, the next step was putting it in a format that would be compatible with the mapping program. My initial plan was to do a map using Google Map Engine, as it has by far the easiest user interface. I intended to do a layer for each year, and size the point markers by the number of troops at that post. However, there is no real good way to make the size of the marker correspond with troop strength, and I would have had to create each point manually, one at a time. For these reasons, I decided to go with Palladio instead. Palladio is a little less user friendly, and is still in development and thus suffers from the occasional bug or idiosyncrasy. However, it does allow you to automatically have your points sized by either frequency or an independent value (troop strength in my case), and you can upload all of your points from a spreadsheet rather than creating them one by one.
While I thus didn’t have to individually create each point, in order for them to populate to Palladio’s map I needed my data to include coordinates for each post. The table in the Annual Report did list each Fort’s location, but only by textual description. Some of these were pretty straight forward, such as “6 miles from Omaha,” while others were more obscure such as “at the fork of the Concho river” or relatively vague such as “in the Owen River Valley.” Obviously this was not going to work for Palladio, so I had to do some additional research first. While time consuming, this was relatively easy. Many of these Forts have become modern towns, and thus show up in a quick Google Maps search (right clicking on the map and choosing “what’s here” will give you the latitude/longitude of any point on Google Maps). I was also pleasantly surprised at the coverage of these Forts in Wikipedia; almost all of these Forts, even those that were only active for a few years and never held more than 100 people, have their own Wikipedia entries. And almost all of those entries contain coordinates, and of those that didn’t most at least included a reference to location tied to a modern town allowing me to Google Map them.
Now armed with coordinates, I built a spreadsheet combining all the data needed for upload to Palladio, specifically Place Name (the title of each post), Coordinates, Strength, and Date (really year, but Palladio requires yyyy-mm-dd format so I used the date of the Annual Report). I did tabs for each year, with a consolidated tab with all years for upload. For the years I chose a representative sampling. While at the archives I pulled the data for all years between 1866 (end of the Civil War) and 1896. However, I quickly decided to start my mapping project in 1876, which marked both the end of Reconstruction (with troops posted throughout the occupied South) and the height of the Indian Wars (the Battle of Little Big Horn, where Custer’s command was wiped out, occurred on June 25/26, 1876). As I began data entry I also realized how time-consuming including every year even during this reduced period would be, and so reduced my table to a sampling of years. I included every 5 years (1876, 1880, 1885, 1890, 1895) for periodic references and then chose several sets of paired years tied to major events. 1876/1877 show the end of Reconstruction, 1880/1881 mark the initial Congressional approval of the Army’s proposal to reduce the number of bases, and 1892/1893 mark what is generally considered the end of the Indian Wars (although clearly no one knew this at the time).
It was at this point in the project that I ran into my first problem. Stupidly, I did not look closely at Palladio before beginning to build my spreadsheet. My coordinates were in degrees, minutes, seconds format (the format that Wikipedia uses) while Palladio requires coordinates to be in decimal format. Rather than back through and manually reenter each coordinate (Google maps gives you both, if you click on the coordinates in Wikipedia it will take you to geohacks, which allows you to see the point using any number of maps, including Google Maps), I decided to attempt to convert them all in excel using a formula. This was harder than I expected at first, and I had to fight through that for a bit. This involved first breaking up each coordinate into individual latitude and longitude columns, putting them in an hh:mm:ss format, multiplying them by 24, and then recombining them into a single column. The end result was that I (finally) had my coordinates in a Palladio-friendly format (although I am still not sure I wouldn’t have been able to do it faster by just reentering them manually).
I now had everything I needed to run the data through Palladio. Here I ran into my second problem. I copied and pasted my data into Palladio, and it showed up perfectly fine on the Data screen. However, when I went to the Map, no points appeared. I tried re-uploading several times, attempted in in three different browsers, and even tried it on a Mac (I’m a PC user as most government systems only work on Windows). Nothing worked. I went back through Palladio’s FAQs and guide for mapping, and I was doing everything right. The points should have been there. Just when I was ready to throw my laptop into the wall, I noticed a single point at the far right of the map (in the Mediterranean). I scrolled the map over, and there were all of my nineteenth-century American army posts….spread across the Middle East and Asia. After a moment of panic that all of my data was clearly wrong, I realized that when I had converted the data into decimal format the E/W had been dropped and all of my Longitudes were positive values when they should have been negatives. After quickly fixing the spreadsheet and re-uploading, all of my forts showed up where they were supposed to be (except Fort Hamilton, which was hanging out off the coast of Spain because I missed a digit in its coordinates…quickly fixed).
The mapping function of Palladio gave a pretty good visualization of the shirinking number of post and the increased size of individual garrisons, but I had to play around with how I used it to get the most value out of it. I had a vision in my head of pressing a play button and watching my map change as posts disappeared and the remaining ones grew larger and larger, but Palladio does not have that level of functionality. The mapping visualization and the change over time visualization are largely divorced from each other; the map shows static data while the Timeline feature allows you to see how various elements of your data changed over time. Hopefully future versions of Palladio will add the ability to have the map show change over time, I can’t imagine my project is the only one that would be useful for.
After using different combinations of data and functions, I decided that the best visualization of my data for my purposes was to actually make use of three separate maps/data sets:
For the first map I included all data for 1876-1896, and sized the points by frequency of appearance of place name. This essentially creates larger dots for the Forts that stayed active longer, with smaller dots for Forts that closed. I also used the Timeline feature, showing how the # of forts decreased over time.
The two other maps are single year maps, one for my starting year of 1876 and one for the final year of 1896. In these maps of have the points sized according to “Strength” column, so that the size of each Fort is represented on the map based on how many soldiers were assigned there. Looked at side by side these two maps give a pretty clear visualization of how the number of Forts decreased while their garrisons increased.
So despite the lack of a timelapse map view, I was able to get out of Palladio essentially what I hoped to achieve. Seeing the data visually, in both the Map and Timeline functions, gives a much more intuitive and emphatic picture of the change in the frontier army over the last part of the nineteenth century. From my start point in 1876 to my end point in 1896 the Army went from 163 posts to 78 in 1896. The size of each post also drastically changed. In 1876 there were only 2 posts with over 500 soldiers and there were 96 (more posts than existed in 1896!) with under 100 soldiers. By 1896, 20 Forts, over 25% of active posts, held over 500 soldiers and only 6 held less than 100 soldiers.
The project also revealed how fluid these posts were, and revealed patterns related to the larger events of the era. Forts would appear, disappear and reappear in response to the campaigns against the Indians or other threats. In 1876 a number of posts still existed throughout the South, enforcing Reconstruction. In 1896 there were additional temporary posts, many containing several hundred troops, in the Northeast in response to labor unrest.
In addition to helping me visualize the changing nature of the Army in the late nineteenth century, I also learned quite a bit about the process of digital history, both in the specifics of Palladio and more generally. First is how much of your time will be spent in the simple drudgery of data entry. I probably spent 80% of my time building my spreadsheet of data and formatting it to be compatible with Palladio; my actual interaction with the digital tools was by far the smallest part of the project. Of the time I spent working with Palladio, probably fully a half was fighting through how to get it to do what I wanted, which left only 10% of my time to actually exploring what Palladio showed me about my data. Still, I was able to get a final product that helped me visualize the data, and even if the results were unsurprising, the visual certainly conveyed it with more emphasis than possible just from looking at the numbers on my spreadsheet. Now I can just hope that the next iteration of Palladio includes a timelapse function for the map.