The Fractal City: What Mathematics Reveals About How Cities Grow

Look at the coastline of Britain on a map. Now zoom in. The jagged edge does not become smoother – it reveals more jaggedness. Zoom in again. More jaggedness still. Benoit Mandelbrot, the mathematician who coined the term “fractal” in 1975, used this observation to upend the way we measure irregular shapes. A coastline, he argued, does not have a definitive length. Its measured length depends on the scale of your ruler. The smaller the ruler, the longer the coast.

Mandelbrot noticed something else: city edges behave the same way.

The boundary between built-up area and countryside, the branching pattern of road networks, the distribution of buildings at different scales – all of these exhibit the self-similar irregularity that defines a fractal. Cities, it turns out, are not merely complicated. They are complex in a precise, mathematical sense. And this complexity is legible in the patterns their streets make when seen from above.

Fractal Dimension: A New Way to Measure Cities

Classical geometry gives us dimensions we intuitively understand: a line is one-dimensional, a plane is two-dimensional, a volume is three-dimensional. Mandelbrot’s fractals occupy the spaces between. A coastline is more than a line but less than a plane – its fractal dimension is somewhere between 1 and 2. The more jagged and space-filling the coastline, the higher its fractal dimension. A perfectly straight coast would be 1.0. The west coast of Norway, with its deep fjords and intricate peninsulas, approaches 1.5.

Michael Batty, professor of planning at University College London and one of the founders of computational urban science, recognized in the 1990s that this framework could be applied to cities. In his landmark 1994 book Fractal Cities: A Geometry of Form and Function (co-authored with Paul Longley), Batty demonstrated that the street networks and built-up areas of cities have measurable fractal dimensions – and that these measurements tell us something profound about how cities grow.

A city that grew organically, through centuries of incremental, unplanned development, tends to have a high fractal dimension: its edges are deeply irregular, its streets branch and curve unpredictably, its density varies in complex ways across scales. London, which grew from a medieval core through centuries of piecemeal expansion, has a fractal dimension around 1.7 to 1.8 when measured by its urban boundary.

A city that was planned on a rigid grid has a lower fractal dimension. Its boundaries are more regular, its internal structure more uniform, its patterns more predictable across scales. The street network of central Manhattan, measured by its space-filling capacity, sits closer to 1.5. The regularity of the grid reduces its mathematical complexity.

This is not a value judgment. High fractal dimension does not mean “better” or “worse.” But it does mean “different” – different in a way that is immediately visible when you compare city maps side by side.

London, Paris, Tokyo: Three Fractal Signatures

Batty’s later work, particularly in The New Science of Cities (2013), expanded fractal analysis to dozens of global cities, revealing that each one has a distinctive fractal signature – a mathematical fingerprint embedded in its street network.

London is the archetypal high-dimension fractal city. Its street map is a tangle of curves, dead ends, irregular intersections, and streets that change names mid-block. There is no dominant grid, no overarching geometric order. The pattern is the product of a thousand years of property boundaries, parish limits, field edges, and speculative developments, each following its own local logic. Seen from above, London’s street network resembles a biological structure – a neural network, a root system, a circulatory map.

Paris presents a fascinating hybrid. Beneath Haussmann’s famous 1850s-1870s boulevards – those dramatic diagonal slashes that give Paris its iconic radial geometry – lies a medieval street pattern almost as tangled as London’s. The fractal dimension of Paris varies dramatically depending on which layer you measure: the Haussmann overlay is relatively low-dimension (those long, straight boulevards are geometrically simple), while the surviving medieval fabric in the Marais or on the Left Bank is high-dimension and intensely complex. Paris is, fractally speaking, a palimpsest.

Tokyo is perhaps the most instructive case. Its street network confounds Western expectations of urban order. There is no citywide grid, few named streets in the European sense, and a block-numbering system that reflects historical land subdivision rather than spatial logic. Yet Tokyo functions with remarkable efficiency. Its fractal dimension is among the highest of any major city, reflecting a pattern of growth that is organic, incremental, and deeply responsive to local conditions. Batty’s analysis suggests that Tokyo’s high fractal dimension is not a sign of chaos but of a different kind of order – one that emerges from the bottom up rather than being imposed from the top down.

What Fractal Analysis Sees in a Map

Here is where fractal geometry connects directly to the visual experience of city maps.

When we look at a map of a planned grid city – Chicago, Barcelona, Buenos Aires – we see a pattern that is relatively uniform at every scale. Zoom in, and the blocks get larger in the frame, but they do not reveal qualitatively new kinds of complexity. The pattern at ten thousand feet looks structurally similar to the pattern at one thousand feet. This is low fractal dimension: scale-invariant regularity.

When we look at a map of an organically grown city – London, Tokyo, the medina of Fez – the experience is fundamentally different. At each scale, new structures appear. Wide arterials give way to medium streets, which give way to narrow lanes, which give way to footpaths, which give way to covered passages. The hierarchy of the network becomes richer as you zoom in. This is high fractal dimension: self-similar complexity.

The visual effect is dramatic and immediate. A grid city, rendered as line art, has a clean, almost mechanical beauty – the beauty of repetition and order. A fractal city has a wild, almost vegetative beauty – the beauty of growth and adaptation. Both are compelling. But they are compelling for entirely different mathematical reasons.

Pierre Frankhauser, a French geographer who has published extensively on fractal urban analysis since the 1990s, has argued that the human eye is naturally attuned to fractal patterns because they mirror the structures found throughout nature – in trees, rivers, mountain ranges, and vascular systems. We may find organic city maps intuitively “richer” because their fractal dimension more closely matches the patterns that our visual cortex evolved to process. A grid, by contrast, has a stark, artificial clarity that appeals to a different aesthetic sensibility – one that values order, transparency, and legibility.

The Map as Fractal Portrait

At streetgrid.art, we work at the intersection of these two aesthetics every day. Every city we render has a fractal signature, and that signature determines the visual character of the finished piece.

A poster map of Paris has a layered complexity that rewards close inspection – the boulevards create a bold macro-pattern, while the medieval streets fill in a dense micro-texture. A poster of Chicago has a hypnotic uniformity, a visual rhythm that extends to the frame edge without interruption. A poster of Tokyo is an intricate, almost calligraphic tangle – a visual experience closer to abstract expressionism than geometric design.

These are not stylistic choices we impose. They are mathematical properties of the cities themselves, made visible through the act of reduction. When you strip away the buildings, the colors, the labels, and leave nothing but the street network, you are looking at each city’s fractal dimension expressed as pure line.

This is what makes map art more than decoration. Each city map is a fractal portrait – a visual record of how a place grew, whether by plan or by accident, whether by the ruler’s edict or by a thousand individual decisions to lay a road, extend a lane, or dead-end a street. The mathematics is in the lines. You just have to look.

Beyond Measurement: Fractals as Urban Philosophy

The fractal perspective challenges one of the oldest assumptions in urban planning: that planned cities are superior to unplanned ones. If organic cities have higher fractal dimensions – if they are, mathematically, more complex – then perhaps their apparent disorder is actually a different kind of intelligence. An intelligence that emerges from the collective decisions of millions of inhabitants, each responding to local conditions, each adding a small piece to a pattern that no single mind designed.

Batty himself has become increasingly skeptical of top-down master planning, arguing in The New Science of Cities that the most resilient urban systems are those that balance planned infrastructure with organic self-organization. The best cities, in this view, are those with moderate fractal dimensions – not so regular that they are inflexible, not so irregular that they are illegible.

This is a mathematical argument for humility in design. And it is visible in every city map ever drawn: the tension between the straight line and the curve, between the planned and the grown, between the grid and the fractal. Two geometries. One city. Every time.