Predicting Air Quality Using Temporal Features: An Analysis using Apache Spark and Machine Learning

Air pollution is a critical environmental issue that significantly impacts human health, climate change, and ecosystems. Developing a highly accurate and efficient air quality prediction model is essential for mitigating these effects. In this study, we utilize the Air Quality Index (AQI) dataset from the United States Environmental Protection Agency (US EPA), focusing on pollutants that pose the highest risk to human health, including ground-level ozone (O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf>), carbon monoxide (CO), and particulate matter with aerodynamic diameters of $\mathbf{1 0} \boldsymbol{\mu} \mathbf{m}$ or less (PM10) and $\mathbf{2. 5} \boldsymbol{\mu} \mathbf{m}$ or less (PM2.5). This paper provides an in-depth analysis of the role of temporal features in air quality prediction. Leveraging Apache Spark and advanced machine learning algorithms, we identify the temporal features with the highest predictive impact using Apache Spark and most promising machine learning algorithms. Moreover, the paper identifies the most promising machine learning algorithm in predicting the air quality.