An IoT-Enabled Intelligent Water Quality Monitoring System for Tourist Safety Using Machine Learning

Recently, Bangladesh has been seen as a popular destination for visitors seeking to see its unparalleled natural beauty, including the world's longest sea beach at Cox's Bazar and the historic Silk Route city of Rajshahi. However, the availability of safe drinking water remains a critical concern, as contamination from industrial discharge, agricultural runoff, and inadequate sanitation infrastructure pose significant health risks to visitors. This study presents an intelligent Internet of Things (IoT) -based water purity monitoring system that combines real-time sensor data with machine learning (ML) for predictive analysis to guarantee safe drinking water in Bangladesh's tourist areas. The proposed system employs four key water quality sensors (pH, turbidity, TDS, and temperature) connected to an ESP32 microcontroller with Wi-Fi capabilities, enabling wireless data transmission to a centralized IoT server. We collected and analysed 3,178 water samples from high-traffic tourist regions, comparing results against WHO and Bangladesh safety standards. Our findings reveal that 51.5% of samples met safety thresholds, while 48.5% were contaminated, emphasizing the need for continuous monitoring. To enhance predictive accuracy, we evaluated five ML models: Decision Tree (DT), Random Forest (RF), Artificial Neural Network (ANN), Bagging Decision Tree (BDT), and Voting Classifier Decision Tree (VCDT). Among these, ANN achieved the highest accuracy (92.66%) in classifying water quality, followed by RF (84.28%) and BDT (83.02%). The system provides real-time alerts to tourists and local authorities when water quality deteriorates, enabling immediate corrective actions.