Design of IoT and ML enabled Framework for Water Quality Monitoring

Globally, the quality of water is declining due to multiple factors such as urbanization, industrialization, agricultural waste disposal, climatic and seasonal variations. As per a study conducted by the World Health Organization (WHO), around 144 million people in the world directly collect and consume untreated surface water from ponds, lakes, rivers, and streams. One key challenge is that the community members are unaware of the dynamic variations in water quality and its impacts on their health, and this demands the need for systems and processes that would provide the opportunity for community members to be aware of these issues. To cater to this demand, this work proposes the design of an IoT system integrated with machine learning algorithms for classifying block panchayat based on water quality index (WQI) and monitoring the quality of water. This will contribute to improving the water resource management and governance by integrating a machine learning approach for predicting location-specific WQI based on current water quality and seasonal factors for the study area. The district of Alappuzha in the state of Kerala is selected for the study as several areas of this district are below sea level and the water sustainability in this area is affected by agricultural practices, tourism, seawater intrusion in summer, and flooding during the monsoon season. Currently, this study area lacks a continuous monitoring system which can help in effective water resource management. Based on the results of statistical analysis, a platform integrated for continuous real-time monitoring of water sources, and mapping of communities into WQI categories has been proposed. In this work, the historic data from 2010 to 2017 is utilized for deriving the community level classification based on the water quality index.