Abstract

Growing energy consumption from campus infrastructure including lecture halls that run heating, ventilation and air conditioning (HVAC) systems motivates data-driven optimization. This research demonstrates an integrated application of Internet of Things (IoT) sensors and cloud-hosted predictive data analytics to enable smart lecture room policies improving efficiency and sustainability. A Raspberry Pi Pico W IoT device was interfaced with BME680 sensor for temperature, humidity and air quality data. The device also incorporated a PIR sensor for occupancy detection and Wi-Fi connectivity to transmit multivariate time series data. The prototype was installed in a university lecture room for real-time data capture. Data was directed to a cloud analytics pipeline including MySQL storage and Node-RED for pre-processing. Time series forecasting was conducted by training autoregressive integrated moving average (ARIMA), Prophet and machine learning models on historical data to predict temperature, occupancy levels, and usage patterns 24 hours into the future. An interactive dashboard visualized both real-time streams and model forecasts using Grafana for analytical insights.