Designing smart Internet of Things (IoT)-driven offices has recently gained considerable attention. Such smart offices are expected to enhance worker productivity and economise electricity for sustainability. However, there are several open issues related to the maintenance of IoT applications. In this study, we implement a calendar-based IoT-driven office exploiting knowledge graphs and rule-based reasoning. The smart IoT-driven office consists of micro-controllers (MCUs) with sensors and actuators, and a controller running on PC. These smart nodes send sensor data which in turn is represented as a Knowledge graph. This has different advantages in terms of interoperability, ease of integration with other systems and improved device compatibility. Decision-making is carried out by a rule-based reasoner. The reasoning process of a rule-based reasoner is more transparent than some machine learning approaches. This allows for easy configuration and adaptation of rules.