Space Syntax and Visibility Analysis

Space syntax began at University College London with Bill Hillier and Julienne Hanson, whose The Social Logic of Space (1984) argued something radical: the pattern of pedestrian movement in a city is predicted less by land use, and more by the pure configuration of the street network - which spaces connect to which. They called the tendency of well-connected streets to attract movement, which then attracts shops, which attracts more movement, natural movement.

The method represents public space not as it looks but as it connects. The oldest representation is the axial map: the smallest set of longest straight lines (sight-and-movement lines) that covers every public space. Each line becomes a node in a graph; lines that cross are connected. Once the city is a graph, you can measure each line's position in the whole - and that position turns out to predict footfall remarkably well.

Two graph measures do most of the work. Integration (mathematically, a normalised closeness) asks: from this line, how few turns does it take to reach all other lines? A highly integrated line is at the heart of the network - easy to get to from everywhere - and these lines reliably carry the most movement; the cluster of them is the integration core of a city. Choice (a normalised betweenness) asks a different question: how often does this line lie on the shortest path between all other pairs of lines? High-choice lines are the through-routes, the ones movement passes along rather than to.

Modern practice has largely shifted from axial lines to segment analysis, which breaks the network at each junction and measures using angular cost (a gentle bend costs less than a sharp turn) rather than counting whole lines. It matches observed movement better. You will still hear "axial"; assume segment-angular unless told otherwise.

The street graph explains movement; a second family of tools explains experience at a point. An isovist is the entire area visible from a given location - stand in a square and the isovist is everything your eye can reach before a wall cuts it off. Its size, shape and how far it runs quantify things designers usually describe with adjectives: enclosure, prospect, exposure, drama.

A narrow lane has a small, linear isovist; a plaza has a large, round one; the moment you emerge from lane into plaza - the isovist suddenly expanding - is the spatial event a good urban room is built around. Because an isovist is just a polygon, a GIS can compute it, measure its area and perimeter, and compare candidate viewpoints objectively. Two shape numbers do much of the descriptive work: a large area signals prospect and openness, while a long, thin isovist (low area for its perimeter) signals a channelled, corridor-like view. Where to set an entrance, where to place a landmark so it is first seen, whether a square feels commanding or merely exposed - each becomes a question you answer by comparing isovists rather than trading adjectives.

Visibility Graph Analysis (VGA), introduced by Turner and colleagues, scales the isovist idea to a whole space. Lay a fine grid over a plaza, station concourse or gallery; for every cell, work out which other cells it can see; join all mutually visible cells into one giant visibility graph. Now run the same graph measures as on streets - visual integration (which spots are visually central), visual connectivity, and so on.

The payoff is a heat-map of a space showing where people will naturally gather, which corners feel overlooked, and where wayfinding will fail. It is the standard tool for reading interiors, campuses, transit halls and public squares, and the free program depthmapX (from UCL) computes both axial/segment analysis and VGA. VGA is where space syntax meets the architect's plan directly.