PERTANIKA JOURNAL OF SCIENCE AND TECHNOLOGY

As the building sector transitions towards sustainability, there has been a growing emphasis and research on the interplay and balance between occupants' well-being and energy consumption. This paper investigated the relationship between indoor environmental quality (IEQ) parameters, energy consumption, and occupancy in a smart office environment equipped with sensor devices in a tropical region using data mining techniques, specifically clustering and association rule mining (ARM). The aim was to detect opportunities for energy savings and IEQ improvements. Our methodology, based on an extensive collection of sensor-based data, relates energy consumption and IEQ parameters to human occupancy and translating these associations into rules. Key findings from the mined association rules included identifying benchmark patterns based on occupancy and detecting anomalies. Anomalous rules highlighted potential inefficiencies, such as high lighting or medium power consumption during periods of very low or no human presence, pointing towards opportunities for energy savings. Rules also revealed situations with high CO2 concentration and warm temperatures associated with medium or high occupancy, suggesting opportunities for IEQ improvement through ventilation optimisation. This study demonstrates the capability of the ARM algorithm to uncover nuanced relationships among occupancy, power consumption, and indoor environmental conditions and provides useful indications towards potential energy savings and improvements in IEQ. It highlights the potential of sensor-collected, data-driven strategies for building operational efficiency and sustainability.