Case study · 03

TerraSphere 3D

What happens when you treat Google's Photorealistic 3D Tiles not as a map, but as a place.

The problem. Google's 3D tiles are extraordinary — photorealistic planetary surfaces at interactive frame rates. But most projects that use them treat the result like a fancy map: a static view, a marker, a popup. The place-ness is wasted.

The approach. I built TerraSphere as a workstation, not a map. Cinematic camera maneuvers that fly you from orbit to street level in one smooth movement. 360-degree drone orbits around specific landmarks. Voyager-style guided expeditions that tell a story as the camera travels. A geodesic measuring tool for actual spatial analysis. The whole interface is oriented around being somewhere, not looking at something.

The hard part wasn't the graphics — Google supplies those. It was choreography. A good camera move has to respect the terrain, anticipate the reveal, and leave the viewer oriented. So I made the camera system adaptive: it slows down, pulls back, and reduces motion when it detects smaller screens or reduced-motion preferences.

The outcome. A tool that people use to explore a place before they go, not just to find it. The Voyager expeditions have become the most-used feature — a reminder that guided narrative beats endless freedom when the goal is to learn something.

What I'd do differently. I'd build the measuring tool first. It's the least glamorous part but the most useful, and it forced me to make architecture decisions that would have been cleaner if I'd made them earlier.

← Back to work→