Case studies should run, not scroll.

Most portfolio case studies are a wall of polished screenshots with captions explaining what you would feel if you could touch them. The reader is asked to trust the narration. I wanted mine to work the other way: judge it hands-on.

Rebuilt in code

So the Sam's Club study renders its aisle scan as an actual three.js scene. The planning screens are real interfaces built with the same component library I'd use in production — you can click the tabs, run the scenario, watch the numbers move. The ShelfSight study embeds the live product itself, full screen if you want it.

This is more work than exporting frames. It's also a different claim. A static case study says: I was there when this was designed. A running one says: I can make this exist. For a designer-PM-builder, the second claim is the whole pitch.

Screenshots ask for trust. Demos earn it.

The rules I kept

Interactive can't mean hostile. The 3D pieces never steal your scroll — the page always wins the wheel, and deep interaction is opt-in, one click away. Everything respects reduced motion. And every animation has a job: showing where attention lands, how a forecast settles, how a loop closes. If it's only decoration, it goes.

Even this blog's hero follows the rule — it's a portrait made of sixteen thousand textured tiles you can tilt and scatter. Not because a blog needs 3D, but because the byline should prove the thesis: I build the things I write about.