Understanding occupant behaviour and the resulting impact on Indoor Environmental Quality (IEQ) is crucial for achieving low-energy use in buildings. However, this is a complex task, due to the stochastic and unpredictable nature of human behaviour. Internet of Things and smart devices have made a larger amount of real-time IEQ data available and data mining techniques offer advantages for knowledge discovery by extracting patterns from extensive raw data more easily than conventional statistical methods do. This study aims at investigating if and how combining some paramount data mining techniques (i.e. clustering and decision trees) can provide useful insights in the context of occupant behaviour and IEQ, focusing on residential buildings. To this end, data on IEQ, outdoor climatic conditions and occupant behaviour, collected in summer 2022 in four flats of two multistorey social housing buildings were analysed. Firstly, a cluster analysis was adopted to extract meaningful daily indoor temperature patterns from 495 time-series data curves. Secondly, a decision tree was computed for classification issues, to improve the interpretability of clustering results and evaluate the connection between dynamic influencing factors (e.g. outdoor conditions, spatial characteristics and human behaviour) and IEQ patterns. The findings suggest that the proposed coupled methodology is suitable for detecting occupant behaviour and IEQ profiles, offering valuable insights for enhancing the comprehension of the influencing factors in residential buildings.