UC Berkeley

Transportation research typically conceptualizes walkability based on access to key destinations and availability of traditional pedestrian infrastructures, such as sidewalks and crosswalks. Urban designers, however, suggest that the designs of streetscapes, the contextual built environments surrounding roadways, play an important role in promoting walking by making it comfortable and enjoyable. Despite the potentially powerful influence of streetscapes, however, there remains a dearth of empirical research unpacking how they are evaluated by pedestrians and impact walking behavior. This dissertation targets this empirical deficit by examining how streetscapes relate to pedestrian preferences. It substantiates how streetscapes are an influential type of pedestrian infrastructure that ought to be more strongly incorporated into walkability planning.The dissertation consists of four core studies that use an array of methods, including integrative literature analysis, interviews, biosensory measures, and route choice modeling, to examine how streetscapes shape the preferences and behaviors of urban walkers. The first study is an integrative review that draws on multidisciplinary literature to investigate the theoretical foundations of walkability and the role of streetscapes compared with more conventional transportation factors. It finds that research in urban design and environmental psychology offers robust theory about how streetscapes affect walkers through psychological mechanisms, such as visual complexity that allows walkers to explore and understand their surroundings. This contrasts with emphases in transportation planning and engineering on physical constraints, such as path connectivity and proximity of destinations. The study shows that there are strong theoretical rationales for how streetscapes influence pedestrians psychologically, encouraging planners to bridge disciplinary traditions and ensure that streetscapes are accounted for in how walkability is understood, analyzed, and planned. The latter three studies investigate the validity of theories about streetscape walkability through a variety of empirical methods. The first uses semi-structured walking interviews along a route in Berkeley, California to investigate how demographically diverse walkers without expertise in planning and design observe streetscapes and respond to them when forming preferences about walking environments. While subjects are universally aware of features contributing to streetscapes, such as buildings, trees, storefronts, and other land uses, they are bifurcated about which types of features they favor. One group consistently prefers built frontages, while another tends to prefer more natural, park-like frontages. Both, however, demonstrate preference for features that are more sensorially engaging, consistent with psychology theory suggesting preference for environments that are explorable and provide information about users’ contexts. By contrast, subjects tend to describe less preferable streetscapes as bland or boring. These results confirm that streetscapes are influential for real-world understandings of walkability and suggest that preferred qualities are largely consistent with theoretical expectations, though they may be provided by both built and natural forms. The next study investigates the efficacy of biosensory measures for evaluating walkers’ emotional responses to streetscapes. Mobile biosensors have the potential to collect large volumes of data about walkers’ experiences as they navigate real-world contexts, but there are substantial hurdles to meaningfully interpreting biosensory measures collected in uncontrolled field environments such as urban streets. These measures may also be inappropriately interpreted as direct indicators of emotions such as stress, although they tend to be theoretically limited to indicating arousal, one component of stress response. This study evaluates how biosensory and other physiological measures, including electrodermal activity (EDA), heart rate variability (HRV), skin temperature, visual glances, and walking speed, correspond with streetscape conditions along the same Berkeley, California walking route, and with the same subjects as in the previous study. Among these measures, EDA demonstrates the greatest consistency across subjects and its levels are strongly correlated with a variety of streetscape features. Indicators of emotions developed by combining EDA with preference ratings suggest that many high-arousal environments are exciting, counter to assumptions in prior research that EDA-based arousal is a direct proxy for stress. These results demonstrate how biosensory measures have potential to elucidate emotional responses to streetscapes, but they must be interpreted cautiously. The final study investigates how streetscape preferences are revealed by the routes walkers choose to take. It uses econometric models to compare more than 650 GPS-tracked walking routes from a travel study in San Francisco against algorithmically derived alternatives. Streetscape characteristics are measured throughout the city using an automated approach based on spatial datasets such as building footprints and high-resolution tree canopy areas, then summarized along observed and alternative routes. Models indicate preference for routes that are more continuously lined with buildings and that have smaller setbacks, more storefronts, and more tree canopy, characteristics that may promote perceptual engagement by providing more complex and interactive visual surroundings. Notably, results also suggest preferred ranges of certain streetscape characteristics within the San Francisco context. Streetscape width, measured between opposing buildings across the street, appears to be most favorable within the moderate range of 15 and 25 meters. Likewise, the cross-sectional ratio of streetscape height to width appears most favorable when greater than 1:2 but less than 3:2, consistent with design theories citing distance limitations to sensory perception and preferences for moderately enclosed spaces. Finally, the magnitudes of relationships between streetscape factors and selected routes were comparable to those for additional route distance and uphill grades, suggesting that streetscapes may affect walkers to degrees that are similar to highly physical factors. These findings provide novel evidence of links between streetscape design and travel behavior that are not only measurable, but substantial. Together, these studies provide a corpus of new evidence that streetscapes influentially shape both perceptions of walkability and the behaviors of real-world walkers. Rather than being ignored by transportation planners, streetscape design ought to be embraced as a form of pedestrian infrastructure that may be incorporated into pedestrian plans as an approach for improving urban walkability.