Daniel PARRA-OVALLE, Geography Department, Autonomous University of Barcelona, Barcelona, Spain, Spain
Oriol MARQUET, Instituto de Ciencia y Tecnología Ambiental ICTA, Autonomous University of Barcelona, Spain
Carme MIRALLES-GUASCH, Geography Department, Autonomous University of Barcelona, Barcelona, Spain, Spain
There is a broad academic agreement that to achieve sustainable urban development it is necessary to promote walking over other modes of transport. Decades of research have shown the significant influence of urban form on pedestrian activity levels. However, a fundamental problem yet to be solved is determining the appropriate contextual units for assessing pedestrian exposure to these contextual influences. For example, the predominant research studying contextual influence at the neighborhood scale defines its boundaries using different levels of administrative areas and different buffer measures. Other studies point out that differences in walking behavior would be explained at the street scale, and use small spatial units, such as block faces. On the other hand, studies based on Space Syntax theories examine the configuration of the urban fabric, operationalized with axial and segmental maps. This study analyzes the spatial autocorrelation of pedestrian volume in order to identify patterns that help define the appropriate geographic scale for the study of contextual influence. Pedestrian volume data are used using an original technique based on the use of an Unmanned Aerial Vehicle (UAV) over two neighborhoods in Santiago, Chile. In addition, autocorrelation is examined in a novel way using neighborhood, connectivity, and direction rules. The results showed some interesting findings. Very high and significant spatial autocorrelation values were found between block faces connected in the same direction, while block faces connected without the same direction obtained very low values. The results suggest that the appropriate contextual units of pedestrian activity should include the directional condition, rather than an isotropic pattern, since it does not act in the same way in all directions. This study presents an original pedestrian observation and analysis technique. These findings are useful for assessing and planning urban influence on pedestrian activity.
Mots clés : walking|UGCoP|UAV|spatial autocorrelation|pedestrian observation
A105238DP