Bechdel Test Visualizer

About The Bechdel Test

The Bechdel Test, sometimes called the Mo Movie Measure or Bechdel Rule is a simple test which names the following three criteria:

It includes at least two named women
who have at least one conversation
about something other than a man or men.

The test was popularized by Alison Bechdel’s comic Dykes to Watch Out For, in a 1985 strip called The Rule. For a nice video introduction to the subject please check out The Bechdel Test for Women in Movies on feministfrequency.com.

Introduction

The Bechdel Test Script Parser was a collaborative digital humanities project between myself and my sister, Laurel Karlsson. It is the product of a shared passion for film, feminism, and the creative potential of technology. By combining the talents and interests of myself, a software engineer, and my sister, an English Scholar, we’ve been able to create an innovative data mining tool for film analysis that we hope to continue to work on and improve. You can check out the project on our website, and the open-source code can be viewed here. This project was born when my sister reached out to me for advice on a few ideas she had been mulling over for a digital humanities project. Hoping to do something related to feminist film analysis, she was feeling very limited by her lack of coding experience and hadn’t been able to find an existing data mining tool to accomplish what she was looking to do. I immediately offered up my coding expertise, looking to gain experience by building a tool from scratch which would accomplish the specific needs of the project. After brainstorming several different project ideas, we settled on the one you see here.

Parsing Method

A Technical Explanation

The Bechdel Test Visualizer is a web application built entirely using JavaScript and Node.js. This tool uses a combination of online databases manual script parsing in order to collect insights about a film. First, the script finds the title of the film, this is generally on the first page, and for the prototype, it is required to be the first line of text in the text file. Once the title of the film has been identified, I make an API call out to myapifilms.com. From here, I can collect general information about the film, such as genre, release year, runtime, etc. But the most important information comes from collecting information about the actors and actresses in the film. Using this API, I can collect a list of all of the actors and actresses, their character’s name in the film, and what gender they are. The character’s gender is the most important bit of information I can collect since it is essential when trying to determine if a scene passes the Bechdel test or not. Once all of the gender and film data has been collected via the API – I can begin analyzing the script. The first step is to break the script down into scenes in order to tell if two females shared a scene together. Scenes were determined by looking for a user-defined set of keywords, such as “int.” or “ext.”. In practice, it was determined that a list of 10-12 keywords well-captured scene changes. In the scene, I collected dialogue by character by using simple parsing techniques. I was then able to test if two female character’s shared a scene. If two or more females shared a scene together, I tested their dialogue by checking if their dialogue contained keywords that indicated that they were speaking about a man or men. This list included words like ‘He’, ‘Him’, ‘Father’, ‘King’, etc. This list is very basic, and after initial tests, I can see that this list needs additional work, as patriarchal words are very ingrained in the English Language, and judging context is very difficult for a computer to do. However, for the context of my experiments, this list has sufficed. This tool, however, is very good at generating data on gender participation in films.