For MS/LIS student Ben Ostermeier, the digital exhibit he curated for the U of I Rare Book and Manuscript Library (RBML), "Starkiller to Skywalker: How Star Wars Evolved from Script to Screen," was a labor of love. A Star Wars fan, Ostermeier spent ten months curating the exhibit, although background work on the exhibit actually started earlier, as a project for one of his MS/LIS courses.
Why did you pick Star Wars for your exhibit?
My work with Star Wars began in my Data Science in the Humanities course I took in Spring 2021 with Professor Ted Underwood. In that class, we analyzed how men and women are portrayed in movies using a data set of more than 600 films. For the course's final project, I decided to focus on the portrayal of gender in Star Wars films throughout their 40-year history, as I am a big fan of the franchise. To do this, I created a data set of the Star Wars original trilogy and Star Wars sequel trilogy. The following summer, I added the prequel trilogy as well.
For my RBML assistantship, I worked on projects to improve our digital exhibits platform and develop a web interface for the digital exhibit site. Since I knew we had the shooting script for Star Wars, and I had already done the work for my class, I decided to create a digital exhibit about the shooting script and use the dialogue data set I already had for the original film as a point of comparison with the script.
How did the library acquire the Star Wars script?
We don't actually know the precise provenance of the Star Wars script, but we know the library acquired it prior to 1982, as the script contains a note that it was transferred to the "Rare Book Room" (as RBML was known at the time) in June 1982. This means it somehow made its way to the University of Illinois Library between the script’s creation in 1976 and 1982.
How did you figure out the word counts for characters and deleted scenes?
For the data set I created for class, I had to go through and identify the speakers of each line of dialogue, which I typically did either by memory or by consulting the film itself or the published script. This was fairly straightforward for most of the dialogue, but Star Wars is rather notorious for giving very minor background characters not only names but extensive backstories.
I digitized the script using our scanner in RBML and then performed optical character recognition on it in the Scholarly Commons, where I am also a graduate assistant. Then I transferred the text into a spreadsheet like the data set for the film itself. To determine word count for both characters and genders, I used the Python Data Analysis Library pandas to analyze the text of both the script and the film.
What was the most interesting fact that you discovered while working on the project?
I already kind of knew this, but the 2017 film Rogue One used archival footage of Red Leader and Gold Leader in its film. This footage was originally shot for Star Wars in the 1970s but ultimately cut from the film, and some of the dialogue for the brief scenes that are in Rogue One is in the shooting script—meaning that George Lucas wrote some dialogue in the 1970s that eventually made its way into a movie in 2017.