How Much Can You Strip from a Map Before It Vanishes?

The art and psychology of minimal cartography — and the threshold where recognition lives

Take a map of London. Remove the labels. Remove the Tube lines. Remove the borough boundaries, the building footprints, the parks, the contour lines. Remove everything except the street network and the river. You are now looking at a tangle of gray lines bisected by a sinuous blue curve. And you still know, instantly and without effort, that this is London.

Now remove the streets too. Keep only the Thames. That single meandering line — its distinctive S-curve through the city center, the sharp southward bend at Limehouse, the broad loops of the Isle of Dogs — is sufficient. One line, and the city is present.

This is the central mystery of minimal cartography: cities possess geometric signatures so distinctive that they survive extreme abstraction. The question is not whether maps can be stripped down. It is how far — and why we can still read what remains.

The Abstraction Scale

In Understanding Comics (1993), the cartoonist and theorist Scott McCloud introduced a concept he called the “abstraction scale” — a continuum running from photorealistic representation at one end to pure iconic abstraction at the other. A photograph of a face sits at one extreme. A smiley face — two dots and a curve — sits at the other. McCloud’s insight was that abstraction does not reduce meaning; it concentrates it. The smiley face is not a degraded version of the photograph. It is a distillation. It strips away everything particular (this person’s nose, these specific eyes) and retains only what is universal (a face, an expression).

Maps operate on exactly the same scale. A satellite photograph of Paris is at the realistic end: every building, every shadow, every tree in the Tuileries is present. A Michelin road map is somewhere in the middle: the city is schematized but still detailed. A Metro map — Harry Beck’s famous 1931 diagram for the London Underground — pushes further toward abstraction, sacrificing geographic accuracy entirely in favor of topological clarity. And a minimal street-network map, the kind we create at streetgrid.art, pushes further still, retaining only the linear geometry of roads and waterways.

At each step along this scale, information is lost. But identity is not. Paris remains Paris at every level of abstraction, because the things that make Paris recognizable are structural, not superficial. The radial explosion of avenues from the Place Charles de Gaulle. The river’s gentle curve. The tight medieval grain of the Marais against the wide Haussmann boulevards. These properties survive any level of simplification because they are properties of the network itself, not of the rendering.

Lynch’s Five Elements

The urban planner Kevin Lynch anticipated this idea in The Image of the City (1960), one of the most influential books in the history of urban design. Lynch asked residents of Boston, Jersey City, and Los Angeles to draw maps of their cities from memory. The results were crude, distorted, and inconsistent in their details — but they shared a remarkable structural coherence. People remembered the same kinds of things, and Lynch categorized these into five elements: paths (the streets and routes people travel along), edges (boundaries like rivers, walls, or highways), districts (recognizable neighborhoods with a consistent character), nodes (focal points like intersections or plazas), and landmarks (distinctive objects used for orientation).

Lynch’s key finding was that these elements are hierarchical. Paths and edges dominate mental maps. Districts are secondary. Landmarks, despite their cultural prominence, are surprisingly minor in spatial memory. People navigate by the linear structure of the city — by its paths and edges — far more than by its objects.

This has profound implications for minimal cartography. When you strip a map down to its street network and waterways, you are retaining precisely the elements that dominate human spatial cognition: paths and edges. You are removing the elements that matter least to spatial recognition: individual landmarks and building footprints. A minimal map is not a degraded map. It is a map that has been reduced to the elements the human brain actually uses to identify a place.

The Signature of the Network

Why can we recognize Manhattan from its outline alone? Why does the curve of the Bosphorus immediately say “Istanbul”? Why is the grid of Barcelona’s Eixample distinguishable from the grid of Chicago, even when both are regular and orthogonal?

The answer lies in what network scientists call the “topology” of the street graph — the particular way in which streets connect, branch, dead-end, and intersect. No two cities have the same topology, because no two cities grew under the same combination of geography, history, planning ideology, and accident. The grid of Barcelona has chamfered corners and a diagonal avenue (the Avinguda Diagonal) that Chicago’s grid lacks. Chicago’s grid is interrupted by the Chicago River and the diagonal slash of Milwaukee Avenue — artifacts of Native American trail paths that predate the grid by centuries. These deviations from pure geometry are what make each grid unique, and they are precisely what survives in a minimal rendering.

The French artist Armelle Caron made this point with devastating clarity in her project Tout bien range (“Everything Tidied Up”), begun in 2005. Caron took figure-ground maps of cities — Paris, Istanbul, New York, others — and then “disassembled” them, extracting every individual city block and rearranging the blocks in tidy rows, sorted by size, like a child organizing Legos. The original maps were immediately recognizable. The sorted versions were not. They were beautiful, strange, and anonymous — generic assemblies of black shapes that could have come from anywhere.

Caron’s experiment proved that urban identity does not reside in the blocks themselves. It resides in the arrangement — in the specific network of streets that separates and connects them. Change the pieces and the city survives. Change the arrangement and the city vanishes. The network is the identity.

The Threshold of Recognition

There is, of course, a limit. Push abstraction far enough and recognition fails. Remove the river from a London map and retain only the street network, and many people can still identify it — but fewer. Reduce the street network to only major roads, and fewer still. At some point, you cross a threshold below which the map becomes generic, and the city disappears.

Where that threshold falls depends on two things: the distinctiveness of the city’s geometry and the viewer’s familiarity with it. A Londoner can identify London from remarkably little information, because every curve of every street is reinforced by years of lived experience. A person who has never visited London needs more — the Thames, probably, and the rough shape of the central area at minimum.

This is why certain cities are more suited to minimal cartography than others. Cities with strong geometric signatures — a distinctive coastline (San Francisco, Istanbul), a dominant river system (Paris, Budapest), an unusual grid pattern (Barcelona, Washington D.C.) — remain recognizable at extreme levels of abstraction. Cities with weaker signatures — sprawling, grid-based metros with no dominant natural feature — require more detail to be identifiable. The threshold of recognition is not uniform; it is a property of the city itself.

Designing at the Threshold

At streetgrid.art, we work deliberately at this threshold. Our goal is to include exactly enough information for a city to be recognizable as itself — and not one line more. This is a design problem, but it is also a perceptual one. It requires understanding what makes each city’s geometry unique and ensuring that those distinctive elements survive the reduction process.

For some cities, this means the street network alone is sufficient. Manhattan’s street grid, with its signature taper toward Lower Manhattan and the diagonal of Broadway, is unmistakable. For others, the waterways are essential — Stockholm, built across fourteen islands, is unrecognizable without its archipelagic coastline. For still others, it is the park system that provides the crucial signature: remove Central Park from a minimal map of New York and the Upper East Side and Upper West Side merge into an undifferentiated field.

This process of selective reduction is not mechanical. It requires judgment about what constitutes the “essential veins” of a particular city — the minimum set of features that preserves identity. It is, in McCloud’s terms, the art of finding the right point on the abstraction scale: abstract enough to be elegant, concrete enough to be recognizable. The stripped-down map that succeeds is the one that triggers an instant, almost involuntary flash of recognition — the feeling of seeing a city you know, rendered in nothing but lines, and understanding that those lines are enough.

That feeling is the entire point. It confirms something we sense but rarely articulate: that the identity of a place is not in its buildings or its landmarks or its labels, but in the deep geometric structure that underlies all of them. Strip everything else away, and the city is still there — reduced to its essential self, and more recognizable than ever.