Product & distribution
I designed and directed a system that turns open map data into a browser world, with AI agents working within the visual rules and checks I set.
Opens in a new tab. Nothing to install.
Best explored on a computer. Watch the video here, or send yourself the link.
Cycling down the high street.
The train comes in.
Dusk to night, time compressed.
From above: the wireframe, then the lit neighbourhood.
01 Get around
A transport system you can actually use.
Wait at the stop and the bus waits for you. Climb the stairs and take the front seat upstairs. Change for the train, walk the carriage while it moves, and step off at the next station. Or walk, cycle or drive. Buses and traffic follow the real roads, the sun and moon keep the local clock, and the weather is whatever it's doing there today.
Bus
The bus pulls in and waits while you board.
Train
The next train comes in under the wires.
Cycle
Take a bike down the high street.
Drive
Or take the car.
Three train lines, 13 stations.
Drawn from the world's own track and station data, with the main roads where buses and traffic run.
© OpenStreetMap contributors (ODbL)
Any weather.
Dry, rain, snow and fog, switched through for this clip. In the world it's today's weather.
Any hour.
Dusk to night on the high street, time compressed.
02 From map to world
I designed the building kit and landmark approach; AI agents research and model within those rules, behind checks I set.
Open map and terrain data provide the streets, building outlines and landscape. I set the visual grammar: ordinary buildings assembled from a kit of researched parts, with the landmarks a local would name modelled individually. The aim is recognisable, not a replica. AI agents research streets, model landmarks and carry out much of the build within that system. I choose the places, set the quality bar and automated checks, review the results and decide what ships.
Drawn in your browser from the map dataCaptured in the world
The whole borough, shown as building outlines from OpenStreetMap.
Each one raised to its height: from the map's own tags, or estimated from the building's type where the map has none.
The same neighbourhood in the world, at dusk.
A whole London borough and neighbourhoods in a Brazilian city.
Hackney, London
Captured in the worldFrom street level to the whole borough, at dusk.
A whole borough generated from open data, with its public streets walkable end to end.
Salvador, Brazil
Test build, ships before launchSalvador's seafront, built by the same system.
Three neighbourhoods generated, with Pituba taken to street-level detail.
03 Where it's going Concept
A proposed way to discover real places through the world.
The idea is missions set by local businesses: find a bakery, discover a record shop's listening night, or seek out a café. Exploring the world could become a way to discover independent shops and plan a visit.
This is a concept. Missions aren't in the world yet.
Concept · example mission
Bakery
Be there when the bakery opens
First loaves out at six. The door at the end of the parade.
Example missions, not real offers.
This one started with a map. Yours could start with a product, a building, a story or a place.
Tell me what you're working on, and we'll work out what it could become.
I read every message and reply by email.
Thanks for visiting!
Starting a new project, or need help improving an existing one? Tell me what you’re working on and I’ll see how I can help.
Prefer to talk it through? Book a free 30‑minute call (opens in a new tab)
