Evaluation and Analysis of Goodness of Fit for Water Quality Parameters Using Linear Regression Through the Internet-of-Things-Based Water Quality Monitoring System

Abstract

Freshwater is the planet’s most important natural resource and is prone to pollution, making it necessary for real-time monitoring. The Internet-of-Things (IoT)-enabled water quality monitoring (WQM) system enables real-time monitoring of freshwater resources. The WQM uses physicochemical parameters, such as temperature, pH, dissolved oxygen, electrical conductivity, biochemical oxygen demand, nitrate, and total dissolved solids to control the water quality. The advent of IoT has proven its effectiveness in capturing, studying, and continuously transmitting environmental data in real time. Mineral-rich watersheds experience the exploitation of available resources in and around rivers, leading to urgent real-time monitoring of river water. The operation pollutes the water by mixing different types of toxic waste, namely, urban, industrial, and agricultural, making it unusable for human activities. In India, the traditional method of taking samples from the site, bringing them to the laboratory, and performing the analysis of the samples is in practice, it takes a day or two to get results and it does not happen in real time, causing water-borne diseases among inhabitants of watersheds. This article attempts to assess the water quality of the Ghataprabha river. Water samples are taken from the river via the WQM system from identified sampling points and subjected to linear regression analysis to estimate the relationships and goodness of fit between the parameters. Once the parameter relationship is known, a one-way ANOVA is applied to the water samples and the water quality is analyzed using the ANOVA hypothesis. Additionally, the river data set can be used to train the WQM system.