# The diagram and the street

> Every transit map lies about distance. The question is which lies help a rider, and which ones send them on a twenty-minute walk between two stations that look next to each other.

- URL: https://editorial-example.notra.site/blog/the-diagram-and-the-street
- Published: 2026-10-02
- Authors: Inés Carvalho, Tomasz Wróbel
- Tags: maps, essays

In 1972, New York got a subway map that was beautiful, legible and widely hated. Massimo Vignelli’s diagram drew every line at 45 or 90 degrees, made Central Park a grey square and put the stations wherever the lines needed them to be. Seven years later the city replaced it with a map that looked like the city again.

People still argue about that decision. We argue about it in our own studio about once a month, usually when a client asks for “a map like London’s”.

## What a diagram is for

A transit diagram answers three questions. Which line do I take? Where do I change? How many stops until I get off? It answers them well because it throws away everything else: the curve of the river, the length of a block, the hill between two stations.

That trade works when riders stay underground. It breaks the moment they come up.

> The diagram is a promise about order, not about distance. Riders read it as both.

## How London taught everyone to read one

Harry Beck drew the first London Underground diagram in his spare time in 1931. The Underground’s publicity department thought riders would find it too abstract, and printed a small trial run in 1933. It sold out. Riders didn’t need the curve of the Thames to be accurate. They needed to know that Bank came after Moorgate.

What Beck understood was that a rider underground has no view to check the map against. There are no streets down there, only stations in order. A map that matches that experience feels right, even when it’s wrong about everything above ground.

Ninety years later, most of the diagrams we draw still follow his rules: straight lines, fixed angles, even spacing between stations, interchanges as rings. They work. Riders in Seoul, São Paulo and Stockholm read them without instructions, which is more than you can say for most interfaces.

## Where the lie gets expensive

We tested this on our own regional map last spring. We showed 140 riders a diagram of six interchange areas and asked them to guess the walking time between pairs of stations. On average they were off by four minutes. On two pairs, where the diagram placed stations side by side but the platforms sat on opposite ends of a rail yard, they were off by fourteen.

![Two versions of the same interchange: a clean diagram where two stations touch, and a geographic map where they sit far apart](https://editorial-example.notra.site/blog/images/figures/diagram-vs-street.svg)

*Left: the published diagram, with Harbour and Mill Lane drawn as neighbours. Right: the same area at true scale. The walk is 1.2 km.*

Fourteen minutes is the difference between a connection and a missed train. It’s also how long it takes for a rider to stop trusting the map.

## Three rules we now follow

We haven’t given up on diagrams. We’ve started marking the places where they stop telling the truth.

1. Draw walking links as their own line type

   A dotted connector with the time on it, like `6 min`, beats two stations that touch.

2. Keep the water

   Rivers and coastlines are the landmarks riders check first. We bend them, but we never delete them.

3. Let the centre stretch, keep the edges honest

   Distortion belongs where stations are dense. At the edge of the network, where a stop can be 3 km from the next, geography matters more than neatness.

## What clients ask for

When a transit authority asks for “a map like London’s”, they rarely mean the angles. They mean a map that people trust on the first look and stop thinking about after the second. Trust is the hard part, and it’s lost at the edges: the interchange that turns out to be a ten-minute walk, the station that looked central and sits beside a motorway.

So the first thing we do on a new commission is ride the network with a printed diagram and a stopwatch. Every place where the walk surprises us goes on a list. That list is usually short, between five and fifteen spots in a network of a hundred stations. Fixing those spots is most of the work.

## The Vignelli question

So was New York wrong in 1979? We don’t think so, and we don’t think Vignelli was wrong either. His map was a very good diagram for a city whose riders, more than most, walk between lines. The fix was never to choose one map. It was to admit that a rider needs both: the diagram on the platform, and the street on the way out.

> Every good transit map is two maps folded into one. The craft is in the fold.

That’s the idea behind the interchange insets we started shipping in Meridian Atlas this year. More on how we draw them in a later issue.
