In the past I have spent considerable time figuring out where a Civil War soldier was stationed as his regiment navigated its mission throughout the war. Sometimes I have been able to extract this information from a book and other times I have relied on the National Park Service (NPS) website. The NPS typically lists places a battle unit has been and battles or other action in which it participated. Here is an example of their listing for the 65th Regiment, Indiana Infantry:
As you can imagine getting all of this information into a simple list of locations and actions can take some time. Today when I needed to perform this task for this particular regiment, I decided to use AI to help. So I put the following prompt into Chat GPT:
Hello, you are a civil war historian. your goal is to build a table of the progression of the 65th regiment in the Indiana Infantry. You can extract the location of the regiment from the following excerpt. Create a code table using this information that shows a list of places the regiment was in the order presented. [I copied and pasted the section after SERVICE from the NPS site here].
Below is an image of a portion of the response by ChatGPT to this prompt with which I was super happy. Even though I didn’t tell ChatGPT to extract the event it did. I figured this is the same thing a worker who was told to do a similar task would probably do thinking “Oh, if I leave that out, they’ll make me go back and add it later so I’ll just do it now.”
Having ChatGPT put this data in a code window made it easy to copy and paste into a spreadsheet. With the data in a spreadsheet, I can do further analysis and can also consolidate events or color code the lines to highlight battles. Having the data in a spreadsheet also helps me more quickly make a map which makes it so much easier to understand where the regiment was throughout the war. Usually I use Google Maps to do this. For Google to plot locations, I need a column with either an address or a latitude/longitude for each location. Having a city and state usually works just as well for the address because I only need general locations. So I added states to any places in the location column in my spreadsheet that didn’t have a state listed. Then I imported the spreadsheet into Google Maps. Here’s the initial result – it does need some cleanup as a few of the entries didn’t have a good location. And I still need to check the plotting for accuracy. But at least this gives an idea of the work saved by having ChatGPT do the heavy lifting for me – it took me probably a half hour to do all this. The last map I made for the 65th Regiment in the Ohio Infantry took me days to do almost the same thing.
