Exploring the impact of building characteristics and occupancy on CO2 levels in 943 Switzerland classrooms

Maintaining healthy CO 2 levels and indoor air quality in school environments is determinant to guarantee children and teachers health, and suitable learning conditions. This study presents an up-to-date assessment of real-time CO 2 levels in 943 primary classrooms across the canton of Fribourg, Western Switzerland, collected during winters from 2020 to 2024. Complementary observation measures were captured, including building characteristics and occupancy patterns. Different statistical analyses were carried out, including linear regression, non-parametric tests, conditional inference trees and hierarchical clustering. The combined use of different methods provides a novel data-driven basis for identifying classroom profiles most associated with elevated CO 2 levels. Overall, mean CO 2 concentrations were highest in naturally ventilated classrooms reaching 865 ppm during teaching hours, compared to 680 ppm for mechanical ventilated classrooms. Conditional inference trees and hierarchical clustering allowed to identify ventilation systems and occupancy as primary determinants of CO 2 levels. Policy contexts, including SARS-CoV-2 protective measures and energy-saving restrictions, strongly affected classroom air renewal, highlighting the interaction between occupancy, technical, and regulatory factors. Although many classrooms met recommended air quality standards, elevated CO 2 levels persisted in some, underscoring the need for targeted ventilation strategies. These findings provide actionable insights for policymakers and school administrators to optimize indoor air quality in both existing and future educational buildings.