A good chart is an argument that doesn't have to make itself. The numbers do the work; the design lets them.
Most charts I see don't pass that test. They're either decorative (pretty, but uninformative) or technical (informative, but unreadable). The good ones — the ones that change how you think about something — share a quality that's hard to name but easy to recognize. They feel inevitable. They look like the only possible visualization of the underlying data. They are almost never the first chart someone made.
This series is about how to get there.
Every dataset has a story worth telling, but the story isn't always the one that comes out of a default chart. Sometimes the right view is a small multiple. Sometimes it's a single annotated line. Sometimes it's three charts where you thought one would do. The craft is figuring out what the data is actually trying to say — and then drawing it well enough that someone else can see it without you in the room to narrate.
I've spent years building visualizations for my work to make decisions from large, messy, complicated data. The lesson I keep relearning: clarity is harder than complexity. Anyone can plot a hundred variables; very few people can show you the three that matter.
The audience I'm writing for is anyone who works with data and wants to be better at communicating what they find. You don't need to know D3 or matplotlib. You need to care about whether the person on the other side of the chart actually understands what you're trying to show them.
My goal is to publish one chapter every month. The math will stay light; the design will be the substance.
If I do this well, you'll start seeing charts differently — in the news, in research papers, in dashboards at work. You'll notice the choices that were made, and the ones that weren't. That's the goal. Once you see those choices, you can't unsee them. And then you'll start making better ones yourself.