Monitoring and Prediction Model of Drinking-Water Quality Based on IoT Architecture

Ensuring drinking-water safety is essential to support the realization of SDGs. Drinkable water has distinct characteristics in flavor, odor, and appearance. Poor water infrastructure and treatment affect these characteristics and can pose a risk to human health. This study resulted in the development of an IoT mechanism for detecting water quality which is affordable and can mitigate the health risks posed on humans. Various water parameters, including temperature, acidity or alkalinity, and impurities, were collected by assembling the Arduino in a low-power energy consumption schema. Considering the limited resource, the real-time notification of data monitoring through a mobile messenger complements web monitoring. The retrieved data was monitored via Grafana and utilized using the binary classifiers of machine learning techniques. Enhancing the artificial intelligence framework was conducted to evaluate the best model between the decision and non-decision trees. The ratio encompassed 60% data training: 10% data testing: and 20% data validation. According to the monitoring and prediction results, authorities can control and manage drinking-water supplies that are routinely tested.