Classification and Forecasting Air Pollution Using Naive Bayes and Prophet: A Use Case of Air Quality Index in Jakarta

Air pollution is a major issue in large cities generally caused by various industrial developments, power plants, vehicles, and natural disasters. Air Quality Index (AQI) that represents the level of air pollution is influenced by several pollutants such as Particulate Matter (PM10), (PM2.5), Sulfur Dioxide (SO2), Carbon Monoxide (CO), Ozone (O3), and Nitrogen Oxide (NO2) which are harmful for health. According to a report by IQAir in 2021, Indonesia has the highest air pollution level in Southeast Asia and ranks 17th globally. Therefore, it is important to pay attention to this air pollution issue to avoid health problems. To address this issue, we aim to develop a system for classifying and predicting AQI in Jakarta. To do this, we utilize Naive Bayes (NB) and Prophet algorithms to classify and predict, respectively, AQI in Jakarta. The data used in this study are obtained from the Open Data Portal of the Jakarta Provincial Government (Satu Data Jakarta). Our results show that the NB could achieve an accuracy score of 99% in classifying the AQI, while the Prophet could predict the upcoming AQI with an error percentage value of 26.88%. In particular, our developed system provides AQI information for today, last six days, and the upcoming week. The developed system also includes a simulation page for checking the level of air pollution and suggestions to act accordingly.