The Beauty of Commute Speed Heat Maps

When traffic data becomes art, you see a city’s circulatory system laid bare

Somewhere in the servers of every ride-hailing company, transit agency, and navigation app sits an enormous, continuously updating dataset: the speed at which people move through cities, measured street by street, minute by minute. In a spreadsheet, this data is numbing — millions of rows of coordinates, timestamps, and velocities. But translate those numbers into color and space, and something extraordinary happens. The data becomes a portrait of the city as a living organism, with arteries that pulse, capillaries that constrict, and blockages that reveal structural failures no urban planner could diagnose from a desk.

GPS Traces as Pointillist Art

The data artist Eric Fischer demonstrated this possibility more vividly than anyone before or since. Working at Mapbox in the early 2010s, Fischer gained access to geotagged datasets of staggering scale — millions of GPS traces, billions of Flickr photos, cellular location pings — and transformed them into images of haunting beauty. His “See Something or Say Something” series on Flickr mapped the locations where people took photographs (in blue) versus where they posted tweets (in red) across dozens of cities. The resulting images looked like bioluminescent organisms — dense clusters of activity connected by glowing filaments of movement, surrounded by vast dark zones of inactivity.

What made Fischer’s work revelatory was not the technology but the aesthetic decision. He chose color palettes that evoked deep-sea photography. He rendered data points as individual luminous pixels against black backgrounds, creating a pointillist effect in which the city emerged not from drawn lines but from the accumulation of millions of discrete human actions. The maps were scientifically rigorous and visually ravishing — proof that the right data, rendered with care, does not need artistic embellishment. It is already art.

Fischer’s approach revealed something that traditional cartography obscures: the city is not its streets. The city is the pattern of movement that flows through those streets. A highway that carries 200,000 vehicles per day is not the same object as a highway that carries 200. They are made of the same concrete, but they are functionally, experientially, and visually different things. Only data visualization makes that difference legible.

The Shape a City Actually Is

The most profound data maps are isochrone maps — visualizations that show how far you can travel from a given point in a given amount of time. An isochrone map does not show the geometric shape of a city. It shows the functional shape. And these two shapes are almost never the same.

Consider a conventional map of Paris. The city appears roughly circular, bounded by the Boulevard Peripherique ring road. Now generate a 30-minute isochrone from the Gare du Nord by public transit. The circle vanishes. In its place you see a shape like a deformed starfish — long tendrils extending along Metro and RER lines, with deep indentations between them where transit coverage is sparse. The 12th arrondissement, which sits close to the Gare du Nord on a conventional map, might be 45 minutes away by transit. Aubervilliers, which appears distant, might be 8 minutes away on the right train.

Isochrone maps reveal that the “distance” between two points in a city is not a spatial measurement but a temporal one, and time-distance is wildly unequal. The urban planner and data scientist Geoff Boeing, whose OSMnx Python library has become the standard tool for street network analysis, has written extensively about this inequality. In his research at the University of Southern California, Boeing has demonstrated that the morphology of a street network — its connectivity, its dead-end density, its intersection spacing — directly predicts how efficiently people can move through it. Grid cities with high intersection density (like Portland, Oregon) produce smooth, roughly circular isochrones. You can see this kind of dense connectivity on a street map of London or a poster map of Tokyo. Cul-de-sac suburbs with dendritic, tree-like street networks produce jagged, irregular ones.

When you visualize these differences as heat maps — with fast-moving areas in cool blues and slow, congested areas in angry reds — the result is a kind of thermal portrait of urban health. The cool arteries are where the city’s circulation works. The hot zones are where it is failing.

Velocity as Color Gradient

There is something deeply satisfying about the aesthetics of speed-based heat maps. The standard color ramp — green through yellow to red — maps intuitively onto our experience of movement. We feel the green. We dread the red. When you see an entire metropolitan highway network rendered in these colors during a Monday morning rush hour, you see a drama playing out: the green freeways of 5:00 AM slowly warming to yellow by 6:30, then a wave of red spreading outward from the city center like an infection between 7:00 and 8:30, before slowly receding back to green by 10:00.

This visual rhythm — the daily pulse of congestion expanding and contracting — has been captured beautifully by organizations like Uber Movement (now discontinued but archived), Google’s traffic layer, and academic researchers using probe vehicle data. The resulting animations are mesmerizing. They reveal that congestion does not simply “happen” on individual streets; it propagates through networks as a wave phenomenon, with predictable patterns of reflection and amplification at bottleneck points. Traffic engineers know this. But only visualization makes it visible to the rest of us.

Mapbox, whose rendering technology underlies many of the most sophisticated modern data maps, has developed techniques for displaying speed data that preserve both geographic accuracy and visual clarity. Their approach to layer compositing — rendering road speed data on top of minimal base maps — demonstrates a principle that data visualization shares with cartographic art: clarity requires restraint. The most legible heat maps use the fewest base-map elements. When you strip away the buildings, the labels, the terrain shading, and let the data speak against a minimal network of streets, the result is simultaneously more informative and more beautiful.

The Art of Invisible Infrastructure

At streetgrid.art, we work with the structural skeleton of the city — the street network, the waterways, the parks that define the negative space of urban form. A commute heat map does something complementary: it shows the invisible infrastructure of movement that animates that skeleton. Our maps show the bones. Heat maps show the blood.

Together, these two kinds of visualization capture something that neither achieves alone. The street network tells you what is possible — where you could go. The heat map tells you what is actual — how fast you can get there. A four-lane boulevard might look identical on a street map whether it sits in a congested city center or a quiet outer district. But a heat map reveals their entirely different characters as conduits of human movement.

There is, we think, something honest about data visualization at this scale. A heat map cannot lie about congestion. A GPS trace cannot fake a detour. When millions of individual journeys are aggregated and rendered as color, the result is a kind of democratic portrait — not the city as a planner designed it or an architect imagined it, but the city as its inhabitants actually experience it, one commute at a time.

Toward a Living Cartography

The frontier of this field is temporal. Static maps — even the most beautiful ones — show a city frozen in a single moment. But cities are not static. They breathe on daily cycles, swell on seasonal ones, and evolve on generational timescales. The next generation of urban data visualization will capture this dynamism: maps that show not just where congestion is, but how it moves; not just today’s transit coverage, but how it has expanded or contracted over decades.

When that happens, the boundary between data visualization and art will become even harder to locate. Already, the best commute heat maps hang comfortably on gallery walls. Already, they provoke the same response as a good painting: the feeling of seeing something you have always known but never quite perceived. The city, it turns out, has always been painting its own portrait in the aggregate motion of its inhabitants. Data visualization is simply learning to see it.