009-14

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

Table 4:

Mean values of grid measures and shape measures compared across the sub-samples of biased, unbiased-sparse, unbiased- dense and the whole sample for urban grids and office layouts

For urban grids, the values of Mean Integration for biased, unbiased=sparse and unbiased-dense types are 0.61, 0.79 and 1.44.

Therefore,

# Integration

Slope

equals

(0.79-0.61)/(1.44-0.61)

=

0.18/0.83 = 0.22 (figure 7c). Integration Magnitude, which is measured as the length of the hypotenuse, equals 0.85. The Mean Depth values for biased, unbiased-sparse and unbiased-dense urban types are 12.40, 15.28 and 5.46, thus resulting in a Mean Depth Slope of 0.31 and a Mean Depth Magnitude of 9.72 (figure 7d) .

Comparison between the typology of urban grids and the typology of office layouts

The newly proposed typology of urban grids is compared to the earlier typology of office layouts which was also distinguished according to degrees of Connectivity and Connectivity Skewness (Shpuza 2006, Shpuza and Peponis forthcoming). The office layout typology, shown with dashed line (figure 7b), is overlaid on the urban typology, shown in doted line. Three representatives (Makarksa-36, Piran Portorož Lucija-26 and Messolonghi-46) and an unbiased-dense office layout (Sears40-o43) are used to illustrate each type and are positioned on the scatterplot of Connectivity vs. Connectivity Skewness according to their data points. Both urban grids and office layouts conform to L- shape clusters and show consistent patterns of splitting into three types. The bottom left quadrants of the unbiased-sparse type for offices and cities roughly coincide. However, the biased urban grids extend vertically displaying more Connectivity Skewness, while the unbiased-dense office layouts are stretched horizontally exhibiting more Connectivity. Office layouts and urban grids mainly overlap falling inside a larger composite L-shaped area with a base of Connectivity 2.4 to 6, and a height of Connectivity Skewness 0.75 to 9.4. This constitutes the first fundamental similarity among linear representations of the built environment.

The measures of Connectivity and Connectivity Skewness are not mathematically related. The string of lines connected in pairs has a Mean Depth with a value close to 2, depending on the number of lines. Hence, the scatterplot can extend rightwards from the y-axis of Connectivity at 2 (figure 7b). In order to illustrate the possibility for having various combinations of Connectivity and Connectivity Skewness falling outside the L-shape cluster of built environment cases, two theoretical linear maps are designed and positioned