Air Quality Forecasting in Urban Environments: A Deep Learning Approach

Air pollution has become an important research topic due to its environmental and human health effects. Today, rapid industrialization and urbanization is one of the major factors in the emission of harmful gases, leading to deteriorating air quality. In this study, air quality problems are discussed, and the adverse effects and consequences of pollutants including sulfur dioxide (SO2), nitrogen dioxide (NO2), carbon monoxide (CO), and particulate matter (PM2.5 and PM10) on human health are assessed. In this study, air quality data from Beşiktaş, Istanbul, has been analyzed by using deep learning models based on Convolutional Neural Networks (CNN), Long Short Term Memory (LSTM), and Gated Recurrent Unit (GRU) to predict air pollutant levels and values. The performance of these models is evaluated using metrics such as Mean Square Error (MSE), Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). The study's findings reveal that the presented GRU model provides superior forecast accuracy for pollutants like CO and NO2, while the CNN model performs better for SO2 and O3 forecasts. This study highlights the importance of using advanced deep-learning techniques for air pollution management. It shows the potential of predictive models to contribute to the policy-making process for sustainable development.