Modelling for Walkability: understanding pedestrians’ preferences in Singapore

Abstract

Objectives Walking is the most prevalent mode of transport in urban environment with regards to the number of trips. As walking is also an inherently sustainable mode of transport, many cities started initiatives to improve walkability. To do this effectively, a profound understanding about the pedestrian behaviour and preference is required. Compared to other modes of transport, pedestrians are more exposed and hence sensitive to the quality built environment (Ewing and Handy 2009). Therefore it is important to account for such factors when studying pedestrian behaviour. This paper presents the findings of a recent study conducted in Singapore whose dense and diverse city centre offers interesting environment to study pedestrian behaviour and whose hot and humid tropical climate pose additional challenges for the built environment to support walkability. The paper has two main objectives: First, to elaborate on the design and conduct of a recent survey on pedestrian behaviour that combines revealed and stated preference methods. Key features of the survey include the deployment of a newly developed App to track pedestrian routes and application of illustrations to describe different walking environments in a web-based choice experiment. The second objective is to present the key findings of the choice models and to evaluate the applicability of the different survey methods to understand preferences of pedestrians. Data & Methodology The city centre of Singapore serves as the study area. It is characterised by high-density developments but also features a wide diversity of walking environments including rows of heritage houses, underground walkways and sidewalks along major urban arteries. For the revealed preference survey, interviewers tracked the walks of 1000 pedestrians by recording route waypoints and conducted a short-follow up survey using a newly developed tablet computer application. The observed routes got map-matched to a detailed pedestrian network that included sidewalks, at grade crossings, over-and underpasses, through-buildings links and walkways through parks. Each link in the network is described by a series of physical and design-oriented attributes which were collected in a previous survey: the featured attributes include for example width of walkway, separation from traffic, noise level, protection from sun and rain, facade transparency and presence of greenery. For each observed route, a set of alternative routes is generated. The statistical analysis of the reveal preference experiment employs path size logit models to account for similarities among the route alternatives. The stated preference experiment, which was conducted as a web-based survey, includes generic, hand drawn illustrations of the built environment, information on walking time, the number and type of required crossings as well as time of day and weather conditions as background information. The variable describing the built environment include the walking environment (street, park or underground walkway), availability of activated building frontages (shops), level of greenery, type of the road, separation from traffic and the availability of protection from rain and sun. Expected results The paper will present a short descriptive analysis on the socio-demography of the participants and the observers walks. The key findings of the respective choice experiments will include the valuation of different physical and urban design related factors to pedestrian route choice. Based on this, conclusions will be drawn about the appropriateness of the applied methodologies to understand preferences of pedestrians in Singapore. References Ewing, Reid and Susan Handy (2009). ‘Measuring the Unmeasurable: Urban Design Qualities Related to Walkability’, Journal of Urban Design 14(1): 65–84.