Closely related to isovist analysis is the much broader topic known as space syntax. This research area was developed initially by Hillier et al. (1993) in the context of architectural design and movement behavior in small-scale regions of the built environment, but has been extended to apply to much larger urban areas.

The earliest work in this field focused on so-called axial lines. These lines are defined as the longest visibility lines (or ‘lines of direct movement’) in the 2D public urban space. By analyzing the geometry of streets, open spaces and buildings the 1^st, 2^nd, 3^rd etc. longest lines of visibility can (in principle) be identified and drawn. A set of such lines covering an entire area will contain many lines and many intersections of these lines. Together they define an ‘axial map’. The axial lines with the highest number of intersections are said to have high ‘connectivity’. Figure 6‑25 illustrates these basic elements using part of the streetscape of the town of Gävle, in Sweden. Here the axial lines with the highest connectivity are colored red, with the lowest being colored blue.

Figure 6‑25 Axial lines and connectivity

Source: Axwoman software, Dr B Jiang

A set of additional measures can be computed once a network of axial lines has been constructed. These include examining the hierarchy of axial lines, for example determining the number of other axial lines within a given number of steps (e.g. 3 steps) distance – this is known as the depth of the axial map. However, the computation of axial lines and maps is complex, slow and not an exact process. It also focuses on spatial elements that are not incorporated as standard objects within GIS datasets. For this reason Jiang and Claramunt (2002) proposed identifying ‘characteristic points’ within urban GIS data, and then using the connectivity of these key points to derive similar urban morphology measures to those produced using the space syntax approach. These characteristic points are relatively simple to identify (e.g. road intersections/network nodes, combined with additional vertices included to handle curved routes). Once identified it is then possible to construct a set of urban morphology measures that are strongly correlated with many of those produced using axial lines. The example in Figure 6‑25 was generated using Jiang’s Java‑based axial line software; ArcGIS extensions for urban morphology analysis (Axwoman) and axial line generation (Axialgen) are also available from Prof Jiang.

The concepts behind space syntax and isovist analysis are combined in Turner’s Depthmap software (Turner, 2004). This uses a similar approach to that of Rana (2004b) described in the preceding subsection (Section 6.3.3.1). First, a vector file (e.g. a DXF file) of the building or street layout to be analyzed is loaded. Then a fine grid of points/cells is overlain on the parts of the diagram or map that is to be analyzed. The program computes the Euclidean inter-visibility between each pair of points in the set (essentially a visibility graph) and from this can, for example, color code each cell in a raster view of the map according to the number of other points it is visible from (a form of global visibility ranking). With this starting point, a range of visibility and connectivity measures can be computed. Furthermore, using a similar starting point with a greedy algorithm, Depthmap can also automatically generate an axial map and related measures. Figure 6‑26 provides simple illustration of these concepts. This shows the internal space of an art gallery with its rooms and connecting doors. Locations that have the highest levels of overall visibility are colored red and yellow, and the computed axial lines, with color coding, are also displayed.