Jaywalking Index: Visual and Socio-demographic Patterns in London

Abstract

This quantitative research delves into the intricate dynamics of pedestrian safety, urban design, and behavioral analysis within the overarching framework of Vision Zero principles in London, UK. With a specific emphasis on comprehending jaywalking behavior, this study investigates the sociodemographic characteristics of jaywalkers and examines the correlation between urban design features surrounding jaywalking crashes. Employing GIS, the research analyzes 25,732 pedestrian crashes and utilizes Visual Artificial Intelligence to segment 280,000 images obtained from Google Street View. Key findings encompass the sociodemographic profiles of jaywalkers and the formulation of a jaywalking index, which serves as an initial tool for identifying areas warranting further investigation in urban design. This index aids in pinpointing regions with a heightened probability of pedestrian crashes, offering valuable insights for proactive urban planning and safety enhancement measures.

Keywords Urban Design; Urban Science; Mobility; Visual Artificial Intelligence; Computer Vision.