Indoor environmental quality (IEQ) has an influence on peoples’ health, cognitive performance and productivity in school and office environments. This study used environmental temperature, humidity, air pressure, and CO2 sensor data collected during 3.5 to 7 months in an office and a school facility to classify occupants’ perceptions of IEQ, stress and productivity in two classes, “negative” and “positive”. Self-reported data from 15 office workers and four teachers were used to train person-specific SVM classifier models. Relatively high accuracies were achieved in classifying IEQ (84%), stress (88%) and productivity (92%) using different combinations of environmental sensor data. Furthermore, the associations between the Big Five personality trait variables (neuroticism, extraversion, openness, agreeableness and conscientiousness) and negative experiences regarding stress, productivity and IEQ were investigated. Positive correlation was found between extroversion and co-occurring stress and IEQ problems, which suggest that more extroverted people more likely to be stressed by insufficient environmental quality or to be more sensitive to environmental factors when under stress. Overall, the results indicate that it is possible to measure and classify perceived IEQ, stress and productivity sufficiently accurately using inexpensive environmental sensors.
- Indoor environmental quality
- machine learning