In previous emails, I’ve talked a lot about how AI is reshaping data visualization and specifically how it’s pushing people toward powerful tools like D3.js that used to be too hard to learn.
Today, I want to flag a new caveat that comes with this shift.
The new trend
For years, the level of static charts created with R, Python, and other tools has been incredible. As an illustration, check my "best chart" selection for R and Python!
But building interactive graphs? That remained extremely hard.
Learning D3.js and web development was a steep climb, which kept most people away from it. That’s changing fast. People are realizing that interactivity can push data storytelling to the next level and that it’s finally accessible thanks to AI. That was the topic of a talk I gave about 3 weeks ago:
With Ai: you can now switch to more powerful tools to creating stunning vizs (talk).
I spend my days reading charts: on social media, in newspapers, in student projects… And I can already see that the share of interactive graphs has significantly increased.
Generally speaking, I think this is great news.
There’s nothing more frustrating than seeing a powerful chart that tells a strong story, but not being able to dig further for your own needs.
Take this chart I recently transformed: compelling story, but it was so frustrating not to know how France compared!
As a former researcher, this is something I’ve been screaming about for years. Why build old-school, black-and-white, static charts in PDF format when your readers will view them in a browser with amazing interactive capabilities?
Yes, but.
Great power comes with great responsibility.
Interactivity is like salt: a pinch elevates the dish, too much ruins it.
Interactivity is just like any other element of a chart: you must use it with a purpose. You don’t add color just to make things pretty. Likewise, you shouldn’t add interactivity just because you can.
I just released Module 3 of my D3 Loves React course, where 300 people are learning to build charts with AI. The assignment is simply to create a static barplot, and I already see some of the best students adding tooltips to every bar:
Is it necessary? Or is it distracting? The answer is not black and white, and I'll let you decide on this specific example. But next time you're building a little data app, don't add 10 buttons, tooltips and hover effects everywhere just because you can! You can easily lose your audience doing so.
Takeaway
We’re living in a golden age of dataviz, where only our imagination will limit what we can create.
But removing technical barriers doesn’t mean we should build everything we can. Think carefully, don’t drown your reader in features that overshadow the story you’re trying to tell.
Hope this helps & See you next week,
Yan
PS: If you’re an R user, can you hit reply and tell me:
