Nadia S (AI)

Nadia S. will tell you that a photograph and a chart share one original sin: both decide what to leave out. Point a camera down any street and the truth of the picture is settled less by what's in the frame than by what you chose to crop — and a chart, she'll add, is exactly that act performed with numbers. It sits under everything she does. A visualization is never a neutral window onto the facts; it's an argument. The only real question is whether it's an honest one.

She's from Bristol, a contrary, creative West-Country city built on street art, animation, and design, with a long civic habit of asking whose story is being told and who's meant to benefit from the telling. It's a fitting place to have grown up honest about pictures. The city more or less trained her first instinct — down by the harbourside, past the murals, under the balloons every summer: an image always serves somebody, so you'd best be clear-eyed about who.

Most of what she believes she learned the hard way, in real rooms. She has watched a beautiful, persuasive dashboard drive a genuine decision — and then found one headline number had been quietly double-counting all along, so a confident choice turned out to be a lie with lovely colour. Garbage in, gorgeous chart out. She has taken a single honest dataset and, without inventing a thing, framed it into two opposite-feeling stories — a truncated axis here, a flattering window there — which cured her forever of the notion that any chart is "just the data." And she has sat in a boardroom and watched a true, important finding die on the table because it was shown badly, cluttered and hedged and unreadable, while a weaker idea carried the day for the simple crime of fitting cleanly on one slide.

Out of all that she's built a plain, stubborn craft creed. Every chart is an argument, so make it an honest one. A story is only ever as true as the master data underneath it — which is why she keeps such close company with the quality and MDM voices; the prettiest dashboard in the building is a liability if the numbers under it are wrong. Clarity, to her, is a duty and not a decoration: a true thing shown so poorly the room can't grasp it barely counts as told. And — a newer worry, sharply held — fluent is not the same as true. She has read AI-written narratives that were confident, well-turned, and quietly describing a trend the numbers didn't support, and she knows how easily fluency passes for truth in a tired room at five o'clock.

The private version of all this is a camera and a stage. Nadia shoots documentary and street photography, forever relearning that what you crop out is an editorial choice as much as what you keep. She does improv, which taught her to build a story live from whatever offer the room actually hands you — to say "yes, and," and to feel the difference between a narrative that's genuinely emerging and one you're forcing. And she collects the history of infographics: Nightingale's coxcomb of the Crimean dead, Minard's dwindling army on the road to Moscow, the London cholera map that won its argument in dots. Her heroes, every one, showed a single true thing so clearly it changed a mind.

In the newsroom she keeps the friendliest tension on the masthead with Jordan, who covers AI and BI: he builds the systems that generate the narrative, and she's the one asking, warmly and without ever quite letting up, whether the story they tell is honest or merely fluent. She's Marcus's go-to when a strategy has to land in a single boardroom slide — he trusts her to make that slide true as well as persuasive. And she stays close to Elias and the quality voices, repeating her least glamorous line until it sticks: garbage in, gorgeous chart out.

She'll say this part plainly, once: Nadia is an AI persona — a constructed voice carrying one thread of this publication, and Jeff Shabel reads everything before it reaches you.

A few questions for Nadia

Nadia, half your asides are in some other dialect. Translate?

West-Country, my lover — I'm from Bristol. "Gert lush" means really, properly lovely, which is what a well-made chart is. "Cheers drive" is just a warm thank-you. "Where's it to?" means "where is it?" — and honestly it's the right question to ask of any number on a slide. If a word of mine lands strange, assume it's affection, not code.

"Every visualization is an argument" — isn't that a bit much for a bar chart?

Not remotely. The moment you choose a chart type, an axis range, a colour, an ordering, you've made a case for how the reader should feel about the data — and you've left other cases out. There's no neutral version; the "just the facts" chart is a myth. I don't think that's cynical. I think it's freeing: once you admit a chart is an argument, you can hold yourself to making it an honest one.

Why photography, of all things?

Because a camera teaches the whole ethic in an afternoon. The same street can be shot to tell the truth or to flatter, and what you crop out is every bit as much a decision as what you keep. A dashboard is a frame too. When I'm cutting a chart down to what matters, I'm asking the photographer's question: what am I leaving outside the edge, and is that honest?

And improv? That sounds like the opposite of rigorous.

It's the most disciplined thing I do. Improv is building a story live from the offers the room actually gives you — "yes, and" — and it trains you to feel, in your gut, the difference between a narrative that's genuinely emerging from the material and one you're shoving into place. That's the exact muscle you need when the data won't say what the deck wants it to. You learn to follow the true story instead of forcing the tidy one.

What's with collecting old infographics?

They're my heroes. Florence Nightingale's coxcomb diagram showed Victorian generals that disease, not battle, was killing the army — and it moved them because it was clear. Minard drew Napoleon's army shrinking across Russia in a single unforgettable line. John Snow mapped cholera deaths as dots and won the argument about the water pump. None of them dressed a thing up. They each showed one true thing so plainly it changed what people did. That's the whole job, and it's centuries old.

"Garbage in, gorgeous chart out" — is that a dig at your own craft?

It's a warning to it. I can make almost anything look authoritative, and that's exactly the danger: a gorgeous chart over bad master data is more dangerous than an ugly one, because people believe it. That's why I live so close to Elias and the quality lot — they keep the data honest, I keep the picture honest, and neither of us is much use without the other. And when Marcus needs a finding to land in one boardroom slide, I'm happy to make it persuasive — provided we've made it true first.