Mapping Shakespeare's plays: an experiment

Friends, Romans, Countrymen: lend me your eyes

For the past eight months I have split my time between working at the Folger Shakespeare Library and at Dumbarton Oaks as a Dumbarton Oaks Humanities Fellow. I am fortunate to work as a part of the Digital Media and Publications team here at the Folger, where I have had the opportunity to define my own digital project this year. Within the first few days of starting work my supervisor and I decided that it would be interesting for me to create a geographical visualization of Julius Caesar to map the play. It can be difficult to imagine the scale of Shakespeare’s plays when we read them or see them performed, because our perspective as the audience is limited to the speeches and actions of characters. For example, in the beginning of Antony and Cleopatra Antony’s trip from Alexandria to Rome takes only twenty pages in the Folger Edition of the play, which corresponds to only three scenes on the stage. According to ORBIS, however, this trip would have taken Antony at least sixteen days if he had perfect weather conditions and made no stops:

Marc Antony’s trip from Alexandria to Rome, with the likely route he might have taken.

The product of this initial foray into the world of Julius Caesar is the History Plays Map.

The journey to the map

When I first envisioned a map of Julius Caesar I thought that I would have one or two points that moved around a map of the Mediterranean, following the characters in the play. However, when I started looking at the play, two things made me reconsider this type of visualization. First, the actual setting of the play is very static, so its simulation wouldn’t be very revealing or interesting. I also noticed that, despite the fact that the setting was relatively static, the characters referred to many Mediterranean locations throughout their dialogues. I decided that the most interesting visualization would be to map all of the toponyms (place names) that appear anywhere in the play.

What’s in a name? Or a toponym…

In my first attempt to extract the toponyms from Julius Caesar I used the XML version of the file from Folger Digital Texts and I wrote a script in Python to extract them as efficiently as possible. My algorithm for finding the toponyms was very simple. First, I extracted all capitalized words from the XML file, preserving their respective Play, Speaker, and Act.Scene.Line information as metadata. Next, I removed all of the stop words (words like ‘the’ and ‘a’) and character names from this collection to narrow it down. Then, I created a list of all the unique terms I had left. I reviewed each term and removed all those which were not toponyms. Once I had a list of all the toponyms in Julius Caesar, I manually found the geographic coordinates for each place and associated them with the appropriate name.

Locating the names on a map was first step in the process where I had to make an executive decision in order to finish the project in a reasonable amount of time. I had neither the expertise nor the time to figure out Shakespeare’s intentions for each toponym, so I tried to get the most historically accurate coordinates from sites like DARMC and Pleiades, and when that wasn’t possible I used the coordinates of the modern location from Google Maps.

Geography in context

Once I had the coordinate information for all of the toponyms, I wrote a script that combined the coordinate data with the record of each toponym and its context in the play and then transformed it all into GeoJSON. The GeoJSON file was all I needed to form a layer for a map in many different mapping softwares. I decided to use MapBox JS for the History Plays Map, because I was already familiar with JavaScript and I wanted to make sure that my map could run on anyone’s browser.

I got to the point of mapping the toponyms in Julius Caesar quickly enough that I realized I could use the same algorithm and scripts to create the same type of data for all of the history plays. The only part that I changed about my algorithm was that I used the TEI Simple versions of the texts (rather than the XML I initially used) because they had all of the proper nouns tagged, which helped me narrow down my search for toponyms much more efficiently.

Finding your way in the map

Finally, the part of this process that took the longest time was actually creating the interactive map. My first issue was that it didn’t make sense to map all of this data at once, but I wanted to make sure that users could find the places they wanted to find. To solve this issue I created the selection menu, which is the first page that users see when they open the map. I also wanted to make sure to not separate my data from the texts too much, so I made all of the line number codes into links to that position in the text on Folger Digital Texts. In the earlier versions there was a feature that allowed a user to click on a character’s name and then only that character’s lines would be mapped. After some rounds of user testing, however, I decided that it was not useful enough to justify the confusion it added to the experience.

The last feature that I added was the “Historical Maps” dropdown menu. I got high quality TIFF images of historical maps from the Folger’s collection and georeferenced them using QGIS. You can use the “Historical Maps” dropdown to overlay these maps on top of the base map like so:

This project expanded and changed more than I had anticipated and I am grateful for the resources and support I had access to at the Folger. I am also fortunate that I had the time to experiment with the data and complete this project. I hope that the History Plays Map will be useful and thought provoking for a variety of audiences.