Anomaly detection analysis for water quality data : A Case Study of Varthur Lake, Bangalore

Water, a crucial natural resource for all living organisms, requires continuous monitoring to ensure safety during utilization. The escalating impacts of urbanization and industrialization have led to a decline in water quality in urban reservoirs such as lakes and tanks. This deterioration is attributed to heightened anthropogenic activities, including illegal discharge of sewage and industrial effluent, inadequate sanitation, exposure of river sites, and urban runoff. Consequently, various urban water resources experience a decline in water quality. Anomalies in water quality data can significantly impede subsequent analyses, resulting in suboptimal decisions or erroneous conclusions. This paper addresses the challenge of anomaly detection in water quality data through machine learning approaches. The study focuses on the Vartur Lake, a surface water body in Bangalore, conducting a vital step in water quality analysis. Multiple anomaly detection techniques, namely local outlier factor, isolation forest, One Class SVM, and Elliptic Envelope, are applied to the dataset. To evaluate the results of the anomaly detection algorithms, a set of algorithmic steps is proposed and implemented for the Anomaly Analysis of Water Quality Data from Vartur Lake in Bangalore.