From Association to Causal Screening: Adjusted Effects of High-Flow and High-Wind Conditions on Pedestrian Emotional Responses in Urban Walking Environment

Urban walking spaces are shaped by both relatively stable visual–spatial features and dynamic contextual exposures, yet the independent emotional effects of these exposures remain unclear. This study develops a causal screening framework to examine the adjusted effects of high pedestrian flow and high background wind speed on pedestrian emotion. Using 9552 participant-location observations collected along a 1.0 km walking route in Changsha, China, pedestrian emotion was expressed as a baseline-referenced emotional response (BRER). Propensity-score trimming established treatment-specific overlap before applying partially linear regression, double machine learning, and causal forest estimation. High flow showed a stable negative adjusted effect with broader empirical support and some heterogeneity, whereas high wind also showed a negative estimate but was treated as exploratory because of limited independent exposure support. Overall, dynamic social and meteorological exposures may retain adjusted effects on pedestrian emotion beyond the observed visual–spatial environment, with stronger evidence for pedestrian flow. Urban walking-space design should therefore strengthen its capacity to buffer transient social and meteorological pressures.