International Journal of Scientific Research and Engineering Development

The degradation of air quality has become a significant challenge impacting environmental sustainability and public health in regions undergoing rapid urbanization. Traditional monitoring systems, which depend on a small number of fixed regulatory stations, often miss localized changes in the levels of pollutants. To get around this problem, modern smart cities need air quality monitoring solutions that can be scaled up, work in real time, and are smart. Recent improvements in machine learning, deep learning, the Internet of Things (IoT), and data analytics have made it possible to make air quality management systems that can predict and change. This study presents a holistic framework that amalgamates sensor calibration, spatiotemporal deep learning models, and explainable artificial intelligence (XAI) to enhance the precision of air quality forecasting. The framework uses multidimensional time-series data, such as weather, traffic, and pollution data, to make predictions more accurate. A comparative analysis of models including ARIMA, LSTM, Transformer, and Graph Neural Networks (GNNs) illustrates that deep learning techniques substantially exceed conventional statistical methods in elucidating intricate dependencies within air pollution data. SHAP-based explainability also helps us understand model predictions and find the most important things that affect PM2.5 levels. The proposed framework shows how AI-driven solutions could make smart cities better places to live by improving environmental monitoring, helping people make better decisions, and improving their quality of life.

Shifa Bilal Tamboli, Simeen Phiroj Mulani, Arman Tajuddin Shiakh,"Machine Learning and Data Analytics Framework for Smart City Air Quality Monitoring" International Journal of Scientific Research and Engineering Development, V9(2): Page(2524-2533) Mar-Apr 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.