Water Quality Monitoring on Streaming Data

Abstract

The increasing contamination of natural water bodies due to diverse human activities necessitates a comprehensive approach to monitoring water quality, especially considering its widespread use in daily life. This study addresses the escalating contamination of natural water bodies, emphasizing the need for a robust real-time water quality monitoring system. Focused on evaluating Triveni Sangam, Prayagraj, where the Ganga and Yamuna rivers converge, the study recognizes the crucial role of continuous monitoring in safeguarding precious water resources. To achieve this, a sophisticated framework has been proposed, leveraging a Spark server to simulate streaming data. This dynamic approach ensures uninterrupted and real-time assessment of water quality, crucial for the effective management of water resources. The system categorizes training data using the Water Quality Index (WQI) and employs Naive Bayes classification for real-time data, achieving an impressive accuracy of 82.21%. The results underscore the effectiveness of learning from streaming data, emphasizing its utility for monitoring water quality in real-time. This study contributes significantly to ongoing water resource management initiatives but also highlights the pivotal role of machine learning in addressing pressing environmental challenges.