S1 File - <p> </p>

Table S1. Summary of industrial parks (n = 31) with details of location, name, area, coordinate, and number of subjects who exposed facial expressions in North China. Table S2. Analysis of variance (ANOVA) of industrial park location (Park, P), season of investigation (Season, S), and their interaction (P × S), as well as gender of subjects and year of facial photo collections on facial expression scores for happy, sad, and neutral emotions of subjects who exposed facial photos in industrial parks of North China. Figure S1. Process of street view image analysis for quantifying green view index through a pre-trained deep learning model. Figure S2. Dynamic changes of facial expression scores for happy (A), sad (B), neutral (C) emotions and PRI scores (D) across four seasons. Error bars mark standard errors. Different lower-case letters show significant difference according to Duncan test at 0.05 level. Figure S3. Whisky-box plots of facial expression scores for happy (A), sad (B), neutral (C) emotions and PRI scores (D) between female and male employees in IPs at regions of North China. Hollow dots mark 5% (top) and 95% quantiles (bottom) and the dash line indicates mean values. Different letters indicate significant difference between genders according to Duncan test at 0.05 level. Figure S4. Nomograph revealing relationships between facial expression scores for happy, sad, neutral emotions and PRI scores. Full lines in red color indicate positive relationships and those in blue indicate negative relationships. (RAR)