Thirteenth International Conference on Design Principles & Practices

Abstract

There is an upsurge in quantitative studies that attempt to correlate urban design concepts with measurable variables in the urban environment. At the same time, enormous advancements have been made in the software that urban designers use to create plans and models. This has substantial implications for the longstanding debates on (a) the role of the designer/planner as an expert serving the public good or elite interests; and (b) the concern of urban design as being a process of refinement [of ideas], rather than a set of specific, memorized, variables that are translated into a physical product. This paper assesses the extent to which a model (by Ewing et al, 2009) for measuring walkability on residential streets may be used to simulate walkable environments on Google Sketchup. Ewing’s model determines 5 urban design qualities that are associated with walkability and it correlates each concept to about 6 variables. My study uses each variable to produce three scenes: (1) where the variable is too high; (2) where it is too low; (3) and where it is at the rate suggested in Ewing’s model. These scenes are then assessed by a panel of urban design experts for the presence of the five urban design qualities. The study finds that the statistical significance of the variables in Ewing’s model is significantly reduced (for all but two of the variables) once they are expressed as a simulation, possibly due to differences in details (colour and shadows) between the simulations and real-world environments.