Shpuza; Urban Shapes and Urban Grids: A Comparative Study of Adriatic and Ionian Coastal Cities

1999). Therefore, the relationship between Connectivity and Connectivity Skewness is scrutinized in detail. The scatterplot between the two measures for 50 urban grids, despite a weak correlation, shows that all data points fall inside an L-shaped cluster (figure 7a), which is almost identical to the pattern discovered for linear maps of office layouts (Shpuza 2006, Shpuza and Peponis forthcoming). The statistical splitting of the cluster into three groups is unequivocal and produces three types of urban grids located in three quadrants: First, “biased” includes 12 cases of grids with high Connectivity Skewness and low Connectivity located in the top left quadrant. This type distinctly exhibits cities with long boulevards that cross a large number of secondary streets (e.g. Messina, Durrës, Makarska). Second, “unbiased-sparse” includes 27 cases of grids with low Connectivity Skewness and low Connectivity located in the bottom left quadrant. The lower left corner of this type contains very sparse grids composed of serpentine roads of small towns located in hilly terrains (e.g. Taormina, Kotor, Piran Portorož Lucija). Moving towards the center of the scatter, still inside the bottom left quadrant, one finds larger cities that exhibit a combination of orthogonal grids and serpentine peripheral additions (e.g. Patras, Pula, Syracuse). Third, “unbiased-dense” includes 9 cases with low Connectivity Skewness and high Connectivity located in the bottom right quadrant. The mid- size cities of this type are characteristically composed of simple orthogonal grids that extend almost evenly in both directions (e.g. Portopalo, Messolonghi, Loutraki).

The mean values of the syntactic measures for urban grids, shapes and shape hulls and their skewness are calculated and compared among three urban types and the sample as a whole and are shown in the upper rows of table 4. In comparison to the other two types, the unbiased-dense urban grids have the highest Line Length at 166m and Connectivity at 4.257. The Integration for the unbiased-sparse at 0.61 is the lowest of the three types; it increases at 0.79 for the biased; and increases further at 1.44 for the unbiased-dense. Integration for urban grids increases both due to the increase of Connectivity Skewness, moving vertically from the unbiased-sparse to the biased; and due to the increase of Connectivity, moving horizontally from the unbiased-sparse to the unbiased-dense. However, the increase of Integration due to Connectivity is much more pronounced than the increase due to Connectivity Skewness. Two new measures are proposed to express the degree of change of a measure among the three types of axial maps.

# Slope

measures the relative change of a measure among three

types and equals the slope of the vector defined by the three values of

the measure:

α=

Δy Δx

(2)

The rise

y is given by the difference between values of the measure

for

biased

and

unbiased-sparse

types.

The

run

x

is

given

by

the

difference between values for unbiased-dense and unbiased-sparse.

# Magnitude

measures the absolute change of a measure among

three types and is calculated by the length of the vector defined by the

three values of the measure:

δ=

Δx ^{2 }

+

Δy

^{2 }

(3)

The rise and the run of the vector are defined as in the formula (2) above.

# Proceedings, 6^{th }International Space Syntax Symposium, stanbul, 2007

009-13