Exposure Window Size Exploration: Linking Indoor Air Quality to Fatigue in Office Environments

Carbon dioxide (CO2), total volatile organic compound (TVOC), and particulate matter 2.5 (PM2.5) are common indoor air quality (IAQ) factors associated with fatigue symptoms in offices, which affect workers’ well-being and productivity. A longitudinal understanding of these IAQs’ effects on fatigue remains limited. This study assesses the viability of predicting fatigue symptom in office based on exposures to CO2, TVOC, and PM2.5 over varying timeframes. Indoor air data were collected from one male office worker in his office over 4 months, alongside Ecological Momentary Assessments (EMAs) to evaluate if he experienced fatigue or tiredness in the past hour. Supervised machine learning models XGBoost were trained to classify if the participant would experience fatigue symptoms based on the office air data, with SHAP analysis further identifying significant IAQ factors contributing to these symptoms. The study offers preliminary insights into how indoor air pollutant exposures drive office worker experiences of fatigue from ML perspective.