Agricultural Indicators as Predictors of Annual Water Quality: An Analysis of Interconnectedness and Prediction Using Machine Learning

Water, the chief constituent of ecosystems, critically influences key facets of civilizations like urbanization and agriculture. Its quality, often impacted by agricultural practices such as pesticide use, plays a central role in sustainable development. Recognizing the intertwined nature of water quality and agriculture, this study’s mission is to predict a country's aggregated annual water quality using agricultural indicators. Focusing on the annual volume of pesticide sales and the agriculturally used area per country as primary indicators, the research aimed to provide a streamlined perspective on agriculture's effect on water quality. Analyzing global water quality through the Water Quality Index presented a macroscopic view, categorizing countries based on their mean scores. Through detailed data analysis, relationships between agricultural variables and water quality metrics were established. For the prediction, multiple machine learning techniques were tested and k-Nearest Neighbor was chosen for the best performance. The predictive model was successful in estimating the Water Quality Index with commendable accuracy.