Decoding the impact of audiovisual street environment features on cycling volumes: Insights from street view imagery and machine learning

Abstract

Understanding the relationship between the audiovisual environment of streets and cycling behavior is crucial for designing more inclusive and responsive transport environments. However, most previous studies have concentrated on the macro-level aspects of the built environment, such as land-use mix, while overlooking the street-space quality characteristics that influence cyclists’ on-site perceptions. Although research has explored the role of objective visual features, studies on the audiovisual environment of streets and its perceived impact remain scarce. This study combines street view imagery, Strava Heatmap data, and interpretable machine learning to investigate the nonlinear and interactive effects of street-level spatial quality on cycling volumes. The results indicate that (1) auditory characteristics—especially noise intensity and sound quality—exert a significant and often stronger influence on cycling volume than visual features alone, challenging visual-centric planning assumptions; (2) several spatial features show nonlinear or threshold effects (e.g., green view index and soundscape eventfulness exhibit inverted U-shaped curves), indicating that excessive environmental richness may reduce cycling appeal (3) synergistic and antagonistic interactions exist between visual and auditory elements—for instance, natural sounds can mitigate the negative impact of high noise, while enclosed visual environments amplify it. These findings provide empirical evidence that optimizing—rather than maximizing—street-level sensory stimuli is essential for promoting active travel. Building on this insight, our study offers practical guidance for policymakers and urban designers to implement more targeted and context-sensitive interventions, and underscores the potential of integrating audiovisual design into sustainable transport planning.