Topological clustering in investigating spatial patterns of particulate matter between air quality monitoring stations in malaysia

Abstract

Air pollution is a persistent issue that arises worldwide. Mitigating this issue poses a significant challenge due to endless industrialization activities, rising construction works and vehicle emissions on the road. Particulate matter (PM₁₀) with an aerodynamic diameter of less than 10 µm is one of the primary pollutants that significantly impact human health. Identifying the spatial pattern of PM₁₀ yields spatially homogenous areas and reveals the most polluted sites and potential PM₁₀ sources of pollution. Typically, traditional clustering is used to achieve the objective. This study proposed a different clustering approach by opting topological data analysis (TDA) technique, namely the Mapper algorithm, to investigate for the spatial patterns of PM₁₀ behavior between 52 air quality monitoring stations in Malaysia. A topological clustering approach produces a topological graph revealing the homogenous region with similar PM₁₀ behavior and a more profound understanding of geographical relationship is achieved through the graph. The topological graph depicts better visualization of cluster similarities pertaining to the study areas compared to traditional clustering, highlighting the advantage of such an approach in investigating the spatial patterns of air pollutants.