A data-driven framework for predictive indoor overheating risk assessment and energy-aware building operation using open IEQ sensor data

This study presents a data-driven framework for assessing indoor overheating risk using open indoor environmental quality sensor data. The dataset utilized in this research is obtained from an existing large-scale indoor monitoring study, comprising high-resolution time-series data collected from IoT-based sensing devices installed in residential buildings. The dataset includes measurements of temperature, humidity, pressure, and light intensity recorded at one-minute intervals over an extended period. The dataset was preprocessed, statistical features were extracted, and a thorough overheating risk assessment model was created using Python-based data analysis tools. With almost 76.48% of observations above 26 °C, 48.24% above 28 °C, and 21.57% above 30 °C, the data show that indoor temperatures routinely surpass thermal comfort thresholds, indicating persistent and severe overheating conditions. Overheating rises in late spring and summer, especially between April and August, according to seasonal study, while the danger is largest in the afternoon and evening, according to diurnal trends. Furthermore, feature importance analysis highlights the considerable temporal reliance in indoor thermal behavior by showing that lagged temperature variables are the most significant predictors of future overheating. A composite risk index was created to effectively classify interior settings into Low, Moderate, High, and Severe risk categories by capturing the combined impacts of overheating intensity, duration, and frequency. The results verify that overheating is a persistent, time-dependent problem impacted by architectural features and environmental variability rather than a singular occurrence. The proposed framework has practical applications in predictive HVAC control, cooling energy demand management, smart building operation, and early-warning overheating mitigation systems. By identifying high-risk overheating periods before critical thermal conditions occur, the framework can support energy-efficient cooling strategies, reduce unnecessary HVAC operation, and improve occupant thermal resilience under climate change conditions. Stakeholders can identify high-risk times and areas, apply focused mitigation techniques, and enhance occupant thermal comfort and energy efficiency by incorporating the model into building monitoring systems. This study shows that proactive indoor overheating risk assessment and control may be achieved in a scalable and efficient manner by utilizing available datasets in conjunction with Python-based analytics.