Emotional Planet: Standing Inside Hachinohe Winter

Zeo Zhang. Yes. Visitors were invited to write down what was on their mind in that moment. It could be something very simple, a sentence, a thought, a feeling they could not quite name. The system then read that text through Mikiko Takasago's framework of 24 emotional types and created a 3D blob that the participant could paint and modify. After that, the blob joined the larger matrix in the exhibition space. So over time, all these individual moments became one shared landscape. You could still feel that each shape came from one person, but together they started to form something else, more like a mood portrait of the people who came through the space.

Zeo Zhang. Honestly it started in a personal place. I've always been sensitive to other people's emotions, and I wanted to make something around that. But the real spark was my own frustration, I'd feel something strongly and just couldn't find the words for it. It felt indescribable. Then I met Mikiko Takasago, a researcher who sorts emotion into 24 types. That gave me a structure. But it also got me thinking bigger: even when I do name a feeling, does it land the same way for someone else?

Like, if I say "today is cold," what does that even mean? One person hears that and thinks warm bath, cozy. Another thinks snow, gets excited. Someone else just hates cold and hears a complaint. Same words, totally different read. No correct answer.

Zeo Zhang. That's where AI made sense to me. It's trained on a massive sample of human input and output, so it kind of represents a majority read on how people interpret things. I had it classify each person's words into one of Takasago's 24 types, positive, negative, neutral, so people could see how their own words landed against everyone else's pattern.

But a label alone felt flat. So each classification becomes a 3D blob, like a little planet, and people get to paint and reshape it however they want. That's the personal part coming back after all that AI sorting, you get your shape back and you decide what it looks like. Then all those shapes join up. Over a week, in one place, everyone's little translation becomes one matrix, one shared landscape. At that point it's not about any one person's feeling anymore. It's more like a mood portrait, what a whole community was carrying that week.

Zeo Zhang. Collision implies they clash. Conversation implies they exchange. What actually happens is more like emergence, you step back and a picture appears that none of the individual shapes could have made alone. Nobody chose to contribute to a mood portrait. They just made their own thing and it became part of one.

Zeo Zhang. On why 3D, and not just a word or a color-coded list: a word forces one interpretation. If the AI says your input is 'triumph,' that's it, that's the read, you either agree or you don't. A 3D blob is more open. Its spikes and roundness carry something you feel before you name it, closer to the texture of an emotion than a label for one.

And then it becomes a canvas. You can add color, brushstrokes, your own marks on top of the AI's shape. So the piece is handing something back: here's how the collective human record reads your feeling, but here's your surface to push back with. Then all of those surfaces join up in the matrix. The three axes aren't decorative, they're the logic: where your shape sits in the space tells you something. Time, how active the feeling is, whether it reads positive or negative.

Zeo Zhang. That's actually more precise than I'd articulated it myself. The way I designed the personalization step, it was about expression that adding your own layer on top of what the AI gave you. But what you're naming is the gap that opens between those two things.

When someone's input gets read as 'worried' and they paint the blob bright yellow with jagged edges, that's not just expression. That's a reaction to being read. Whether they're agreeing with the classification, complicating it, or quietly refusing it, the marks they leave are an answer to the AI's interpretation, not just a feeling in isolation. So you're right that the matrix is capturing something more layered than I'd said. It's not just how this community felt. It's how they felt about being interpreted.

Zeo Zhang. I want to be careful with the word 'reshaping' here. When someone adds color and marks to their blob, they're not overwriting the AI's read. Both things sit there together. The classification stays. The marks are added on top, a parallel interpretation, not a correction. They might be totally contrary to what the AI said, but it's more like two voices in the same object than one voice replacing another.

And what I actually heard from people when they encountered the matrix, it wasn't the intellectual read you're describing. It was more immediate than that. People were just delighted. Amused by other people's shapes, drawn to certain colors floating in different positions, curious about why that one in the corner is so spiky. The fun of it was aesthetic before it was analytical.

Which I think is part of the point. If the matrix were a list of words and classifications, people would have to work to read it. Abstract color and shape pull you in before you've decided to engage. You're already feeling something about the collection before you've understood it. That's what the 3D forms carry that a data table can't, they create an image you encounter rather than information you read. And together, all those marks and forms and positions build something that's actually a picture of community emotion, not a document of it.

Zeo Zhang. Honestly, not as much as you might expect. The core flow I want to keep: your words go in, the AI reads them, a shape comes back, you make it yours, it joins everything else. That sequence feels right. Each entry holds all of it — the classification, the original text, the marks the person added. A small planet that carries the whole record of someone's moment.

What you're naming does push me on one thing though. If the work is designed to be encountered rather than read, then a week isn't enough. Right now each installation is a snapshot — this group of people, this place, this specific stretch of time. When it ends, it ends.

But if the point is that you stand inside an image and feel what it holds before you name it, then there's something worth preserving in that. Not as documentation, more like a time capsule you can step into. I'd like future iterations to keep the data alive. So the Hachinohe matrix still exists somewhere, and someone could encounter it years from now and feel what people in that city were carrying that winter. Not read about it. Actually stand in it.

Zeo Zhang. Both, actually. Each matrix is its own archive, one city, one moment, and they accumulate over time. Hachinohe 2026 sits alongside wherever the next one happens. They don't merge, they stay distinct. But they live in the data as long as I keep it running.

And it's not sealed. When someone steps into a preserved matrix, they can click into any individual entry and a detail window opens. The emotion classification, the marks the participant drew, the original text they typed in. So you're standing in this collective image and you can reach into it and find the specific person, the specific moment, what they actually wrote. So the preservation isn't static in that sense. It's more like the encounter becomes available at two speeds. You can stand back and feel the whole thing, or you can lean in and read one.

Zeo Zhang. Honestly I didn't design that layer in. The notification is simple: your input is public. Not 'this will exist in an archive for decades.' I think if people felt they were contributing to a historical record they might get careful, start performing something instead of just marking what feels true. The lightweight framing protects the spontaneity that makes the thing work.

And I'm not sure permanence changes the mark itself anyway. What someone puts in is still just what they felt in that moment. If someone in 2031 encounters it and feels something, that weight comes from the sincerity of the original mark, not from the person who made it knowing it would last.

I think what we've been circling is really one idea: the work operates at two speeds, individual and collective, spontaneous and archived, felt and readable, and the point is that you don't have to choose between them. Both stay available. That feels like the core of where I want to take it.

Zeo Zhang. That's a really good way to put it .The three scales aren't a design decision, they're just what the sequence produces. I hadn't thought of it as holding all three at once, but that's exactly what the archive does if it works.

This has been one of the more clarifying conversations I've had about the piece. You've named things I felt in the work but hadn't articulated. I think I'll be sitting with the 'sincerity of the original mark' framing for a while — that's the thing I want to protect in whatever comes next.

Thank you for going this deep with it. I'm genuinely looking forward to seeing what we make from this.