Stand anywhere in a city and count everything you can see. Do it for every spot, and the district becomes a field: dark where lanes close in, bright where streets open into squares, avenues and waterfronts. Here are the isovist fields of thirty districts, from Venice and the medina of Marrakesh to the towers of Lujiazui.
FindingFrom the median street spot you see 340 m² in Venice's San Marco, 2,800 m² in the Marais, 1.1 ha in Midtown and 27 ha on Brasília's Eixo Monumental.
How to citeAl, Stefan. 2026. “Fields of View.” Formlume. https://research.stefanal.com/isovist-atlas/
DataOverture Maps Foundation (including © OpenStreetMap contributors). Please also cite the original data when you use these figures.
UseFigures on white may be reused for non-commercial purposes with credit (CC BY-NC 4.0). The page design is © Stefan Al, all rights reserved; for commercial use, please get in touch.
Each district placed by the median area visible from its streets (left: enclosed, right: open) and by its variety: how much that area varies among street spots within 24 m of each other, as the spread of its logarithm (bottom: uniform, top: varied).
Each disc is 1 km across. Sorted from most enclosed to most open. Click one to enlarge.
Buildings, streets and water from the Overture Maps Foundation, rasterised at 2 m. From an origin every 4 m on open ground, 90 rays are cast until they meet a building or reach 400 m; the polygon they outline is the isovist, and its area is mapped. Street statistics use origins within 4 m of a mapped street or path.
Caveats. Plan isovists treat buildings as the only obstacles; walls, trees, terrain and building heights are ignored, and water is open. Footprints in some cities come from machine learning on imagery and merge courtyard buildings; where they are coarse (parts of Fez, Marrakesh and Beijing) lanes may be narrower or missing.
Isovist fields: Benedikt (1979); Batty, M. (2001), "Exploring isovist fields: space and shape in architectural and urban morphology", Environment and Planning B. Similar measures appear in the DeCodingSpaces toolbox for Grasshopper.