Ponente
Descripción
Faculty of Engineering, National University of Asuncion, Paraguay. Email: fernanda.carles@gmail.com
Air pollution poses substantial health risks, leading to millions of deaths annually worldwide. In Asuncion, Paraguay, local activities and seasonal forest fires contribute to elevated levels of pollutants, particularly PM2.5 and PM10. To address this issue, the Faculty of Engineering at the National University of Asunción (UNA) has established a network of air quality monitoring stations, providing data since 2019.
Our research focuses on developing machine learning models to predict Air Quality Index (AQI) levels in Asunción. Among these models, window-based regression trees were trained using air pollution and local weather data, achieving accuracy rates of 91% for 6-hour forecasts and 86% for 12-hour forecasts. These predictive capabilities are crucial for timely public health alerts and decision-making.
To make this research accessible and actionable for the public, we are developing an open-source, user-friendly web application. This platform will offer real-time air quality forecasts, interactive visualizations, and alerts, utilizing data from our monitoring stations, ground weather measurements from the Silvio Pettirossi Airport and pollution data from the US Embassy in Asunción’s monitoring station. By providing transparent and real-time air quality information, the web app aims to empower residents of Asunción to take informed precautions against air pollution.
This initiative represents an advancement in air quality monitoring and public awareness in Paraguay, fostering collaboration between academia, civil society, and the open-source community. Our project highlights the potential of data science to address environmental challenges and improve community well-being.
This work is supported by the CONACYT, the FIUNA and the Mozilla Foundation.
Keywords: Air Quality, Time Series Forecasting, Machine Learning, Data Engineering
