Using Graph Data Science to Assess the Impact of Plastic Pollution on Air Quality in Sri Lanka

Abstract

Plastic pollution and air pollution are two pressing global challenges that have often been studied separately, leaving their interactions underexplored. Mismanaged plastic waste not only contaminates land and water but also degrades air quality through open burning and airborne microplastics. Sri Lanka generates over 1.5 million tons of plastic waste annually yet recycles only ~3%. This low recycling rate contributes to severe pollution from unmanaged plastics. Meanwhile, urban air quality has deteriorated, with smog episodes in Colombo often exceeding an Air Quality Index (AQI) of 150. This study assesses the impact of plastic pollution on air quality in Sri Lanka using a combination of systematic literature review and graph data science techniques. This study applied a PRISMA-based systematic review of global and local studies to evaluate the contribution of plastic waste to air pollution. In parallel, graph data science modeling was applied to connect key entities (plastic waste, burning practices, microplastics, air quality indicators, health outcomes) and to identify critical linkages. The literature review revealed that open burning of plastic waste releases toxic particulate matter and gases (e.g. dioxins, furans), and recent studies have detected microplastics in ambient air, posing inhalation risks. Our graph analysis of Sri Lankan data suggests a strong correlation between increasing plastic waste and rising ambient PM₂.₅ levels, and highlights “open burning” of waste as a central node linking plastic pollution to poor air quality. The integrated graph model identified that reducing open waste burning and plastic leakage could significantly improve air quality. We conclude that plastic pollution is not only a marine or soil issue but a major air quality threat, particularly in developing countries. Mitigating plastic pollution (through better waste management, reduced single-use plastics, and prevention of burning) can yield co-benefits for air quality and public health. The novel application of graph data science in this context demonstrates a powerful approach to uncovering complex relationships in environmental systems, filling a critical research gap and informing more holistic pollution control strategies.