July 22, 2026 — by Lulu Liu in Uncategorized
Image via Strauss Center
When a lot of people think about writing in Computer Science or Data Science, they picture lines of neon-toned code on a dark screen, or a stereotypical nerd who hasn’t showered for a week endlessly pressing buttons on his keyboard. But the discipline communicates far more than that. Most of them are multimodal, meaning they combine multiple modes of communication at once: visual, linguistic, interactive… For this post, I’d like to take a step back and look at how my field communicates with itself and with the world.
Take the Jupyter Notebook. On the surface, it looks like a coding environment, but if you dig a little deeper, you will find something closer to a document, with code cells that process input, markdown cells with headers and additional comments, and inline visualizations all coexisting on the same page. That is not something people can usually accomplish in their coding environment. This kind of notebook can contain the written hypothesis, the code itself, graphs, and anything related of the sort. The different modes come together to build and complete an argument.
Image via Jupyter Notebook
Data visualization platforms like Observable take this further by combining interactive charts, written narrative, and live code in a format designed specifically for public communication. Additionally, data journalism outlets like The Pudding use what they call scrollytelling, words that advance alongside animated visualizations, to make statistical arguments accessible to an audience who find old-school-style academic papers to be boring. The underlying info is the same, but the mode of delivery isn’t, and that affects what the reader sees, understands, and remembers.
Screenshot via The Pudding
Talking about a text that is more familiar to those familiar with CS, even a GitHub repo is a multimodal text when you take a close look. This sample’s README combines plain words and Markdown. The file tree is used to organize information. The commit history is a chronological record that tells the story of how the program was built. The badges represent metadata visually. There are so many modes, and none are used redundantly. Each one does its own job.
Example via React’s README
Some of these multimodal genres may seem “flashy”, but they were largely developed out of practical necessity. Code alone does not explain itself, and plain data does not deliver an argument. Over time, the field learned that communicating findings requires ways to frame them, over text, visuals, and more, to make them navigable. The result is a set of genres that are more sophisticated than they typically get credit for, and students entering this field are going to learn to be comfortable with them.
Leave a Reply