Environmental Datasets for Agege, Shomolu, Badagry, and Ikeja

Today, some municipalities experience notably higher temperatures compared to their surrounding rural environments. These are referred to as Urban Heat Islands (UHIs). This study investigated the cognitive and sleep-related impacts of UHIs across four urban regions in Lagos State, Nigeria: Badagry, Shomolu, Agege, and Ikeja. Three Machine Learning models (Random Forest, XGBoost, and Gradient Boosting) were used to analyse a 1-year region-specific environmental dataset to determine the extent to which prolonged exposure to higher urban temperatures affects sleep quality and daytime alertness. This was correlated with sleep metrics derived from the Pittsburgh Sleep Quality Index (PSQI) and Epworth Sleepiness Scale (ESS). The two scales were used to compute weighted scores and to classify severity levels.The models had correlation coefficients of 0.835, 0.980, and 0.980, and corresponding Mean Squared Errors (MSE) of 0.0027, 0.00031, and 0.0031, respectively, demonstrating high predictive accuracy for PSQI scores. From the studied regions, Shomolu recorded the highest weighted score of 2.30, indicating poorer sleep quality, while Badagry had the lowest score of 2.00. The ESS scores also follow the same pattern. These results reveal important regional disparities in the cognitive impacts of UHIs. Thus, locations with low ESS scores mitigated UHIs through coastal cooling (Badagry), etc. Conversely, regions with higher ESS scores experienced intensified UHIs due to dense and unplanned settlements among others.The use of ML models for analysing sleep patterns in this study provides actionable insights for public health officials and urban planners.