The Universal Cartoon
Using web scraping and natural language processing, I found the words that were most used in New Yorker cartoons. I then drew illustrations from these words to create universal images that work aptly with every caption.
Materials used: Python, TextBlob, Pen, Paper, Ink Wash.
As part of the fall 2018 program at School for Poetic Computation, I studied Code Poetry with Sam Lavigne and learned how to use Python to collect, abstract, and artistically manipulate internet data. Sam taught us that simple algorithms can grant access to otherwise obscured information. In The Universal Cartoon, I use these tools to enter the world of The New Yorker and play with the mechanics of humor behind their iconic cartoons.
In my experience, cartooning can be daunting and unpredictable. As Bob Mankoff, former cartoon editor for the New Yorker, said, “there is no algorithm for humor”. The frustrating randomness of the idea generation process prompted me to turn to the Natural Language Processing libraries that we had explored in Sam’s class. Just as people have proposed universal New Yorker cartoon captions, I wanted to see if I could draw universal images that would be funny when paired with any caption.
An incredible archive of all published New Yorker cartoons luckily already existed at The Cartoon Bank. With a bit of finagling, I was able to collect all captions, keywords, and image descriptions form this website. From there, I used TextBlob to run a text analysis on all my ‘New Yorker words’ and pick out the top nouns, verbs and adjectives that appeared in the text.
Using my list of top parts of speech, and advice from current New Yorker cartoonist, Jason Adam Katzenstein, I drew a few images that could be universally funny with every caption. I worked with the words: cat, couple, dog, mouse, caveman, ape, bed, bar, restaurant, doctor, fish, crab, desert island, say, question, friends, enemies, fancy, drink.
At the SFPC showcase I also asked visitors to participate in a mini Reverse Caption Contest to get a better idea about how people responded to and create humor.
I currently use the project as a prompting tool for my cartooning practice and will publish the New Yorker caption canon for public use in 2019.
View the project: https://kchanba.github.io/newyorkercaptions/