DATA-CENTERED DEVELOPMENT AND PREDICTIVE EVALUATION OF A SUPERVISED MACHINE LEARNING SYSTEM FOR WATER QUALITY ANOMALY DETECTION IN A LAGUNA WATER DISTRICT

This developmental–experimental research aimed to design, develop, and evaluate a data-driven system for automating water quality monitoring and anomaly detection in a water district in Laguna. The existing manual process, which involved logging inspection data across 19 pumping stations, caused delays in detecting anomalies and in meeting compliance with Philippine National Standards for Drinking Water (PNSDW) 2017. The system was built using Agile methodology and user-centered design. Historical inspection logs were digitized and used to train a supervised machine learning model, specifically the Random Forest algorithm, to detect anomalies in pressure, discharge, and chlorine residual levels. It also included a predictive feature for estimating chlorine effectiveness duration. A three-month pilot deployment was conducted to assess the system’s performance. The model achieved over 85% accuracy. Usability evaluation through surveys showed a 95% reduction in reporting time and a 90% user satisfaction rating across Functionality, Reliability, Usability, Efficiency, Maintainability, and Compliance. The system successfully improved monitoring efficiency, supported proactive operations, and enhanced regulatory compliance. This research provides a practical model for digital transformation in local water utilities.