Leveraging IoT and Machine Learning to Optimize Public Transportation and Reduce Carbon Footprint

Public transportation is a key strategy for lowering greenhouse gas emissions, yet inefficiencies in demand prediction and resource allocation reduce its effectiveness. This study presents a data-driven framework that integrates Internet of Things (IoT) behavioral data with machine learning (ML) models to improve transit efficiency and mitigate carbon emissions. The IoT-Carbon Footprint Dataset, containing 10,000 daily activity records on energy use, travel distance, and transport mode, was analyzed using regression, classification, and clustering techniques. Linear Regression showed strong predictive accuracy (R² = 0.854, MAE = 2.96), while Logistic Regression achieved 87.1% accuracy in classifying high- and low-emission groups. K-means clustering, despite low cohesion (Silhouette Score = 0.103), identified general mobility profiles such as habitual car users and public transport commuters. The results demonstrate that simple, interpretable ML models can provide robust insights for evidence-based policy, supporting low-emission mobility and efficient transit planning. This framework illustrates the practical application of IoT data and ML in advancing sustainable urban transportation.