Terroirs

Working at the intersection of design and wine education, I kept seeing the same gap: the visual infrastructure that educators needed to teach geography as a sensory argument simply didn't exist. Every tool treated the map as supplementary or an add-on. I thought it should be the explanation.

Wine education is built around maps, but almost universally flat, static ones. Students are expected to understand regional characteristics, but the tools available give them names and borders, rarely the physical reality behind them.

The existing tools split into passive ones (textbooks, PDFs, flat maps) and active but context-free ones (flashcard apps). Educators had no way to teach geography spatially, and students had no way to test themselves with the terrain actually visible.

The intervention wasn't to build a better flashcard app or a more detailed reference map. It was to make the map itself the testing environment, so that geographic context and active recall happened at the same time. That decision shaped every subsequent design choice.

Mission Control

The student dashboard. Organises study by country and region, tracks progress across sessions. Designed to work alongside a textbook rather than compete with it, the positioning is explicit, stating: "Your textbook has a new best friend."

Flight Decks

The core study mechanic. Each country is broken into numbered flights, each covering a group of appellations. Students identify regions on an unlabelled satellite map before labels appear. The drag-and-drop interaction requires a physical commitment to the answer before confirmation.

Spaced Repetition

Regions a student misidentifies are relisted. Applying this spatially means the repetition reinforces the geography alongside the name.

3D Terrain

Regions can be viewed in satellite 2D or elevated 3D, pulling real elevation data rather than illustrated topography. Real topographic data was a deliberate choice: seeing the topographic variation adds spatial context.

Mapbox was chosen specifically for the quality of terrain rendering. This kept the product accessible on any device without installation, which matters for classroom use. The front-end was built using Claude Code, which compressed the gap between design decision and testable prototype.

Terroirs is still in active development. The core map and flight deck functionality works, and the interaction design and custom flashcard features are currently being made fully responsive across screen sizes. The next step is bringing it into real educator workflows. Finding teachers willing to test it as a classroom companion and feed back on what the product still needs.

If you're a wine educator reading this, I'd like to hear from you.

The clearest tension in building this was resisting comprehensiveness. Early versions reached toward being a full wine reference platform. That scope made the core interaction harder to reach and the product harder to explain. Pulling back to the geography problem, made every other decision easier to evaluate. Does this feature serve that moment? If not, it waits.