Understanding Indoor Environmental Quality (IEQ) in residential buildings is often oversimplified due to the complexity of occupant behaviour in correlation with the variability of contextual factors like building characteristics and outdoor conditions. The Internet of Things has facilitated access to a substantial amount of real-time IEQ data through smart devices. Data mining techniques can support the analysis of such wide available databases. In this sense, clustering and classification algorithms offer advantages in identifying complex relationships and revealing hidden structures. They efficiently extract daily patterns from extensive raw data compared to conventional statistical methods. However, the literature on the application of these processes for IEQ seems to be still limited, mostly focusing on energy consumption. Indeed, this work aims at using these techniques in the context of energy-related behaviours, moving towards multidimensional identification of occupant patterns depending on dynamic factors. To this end, a room-level monitoring campaign was conducted in 10 rooms of 4 flats, chosen from two multistorey residential social housing buildings in Reggio Emilia, Italy. Monitoring took place throughout the entire summer period in 2022, capturing data on IEQ, outdoor climatic conditions, and building occupant behaviour. The analysis first involved extracting meaningful daily indoor temperature patterns through cluster analysis applied to 128 time-series data curves. Then, post-clustering analyses using classification trees were performed to enhance interpretability, examining the connection between influencing factors and IEQ patterns. Findings suggest that the proposed clustering method effectively detects daily patterns in the IEQ domain and that the post-clustering analysis outcomes provide valuable insights for a better understanding of the factors influencing IEQ (e.g., the importance of occupants’ behaviour over other characteristics).