Urban Water Quality Prediction Using Open Data

Abstract

How clean the water is in cities affects our daily lives. In order to prevent pollution and safeguard people's health, predictions of municipal water quality are helpful. Still, predicting a city's or other big metropolitan area's water quality isn't a picnic. There are a lot of factors, such weather, water consumption habits, and land uses, that contribute to the non-linear variation in urban water quality. Our goal in this work is to forecast the water quality at a given station over the next few hours using data-driven methodologies. Current screen terminal data on water top-quality and hydraulics is one of our main sources, but we also use weather, pipeline networks, road network frameworks, and sites of interest (POIs) as city-observed information resources. To begin, we use extensive testing to determine which variables significantly impact urban water quality. The second component of our approach is a mechanism for multi-task, multi-view understanding that we use to integrate datasets from several domains into a single model. We test our algorithms on actual datasets. We have conducted extensive experiments to confirm that our technique is superior to other standards and to show that our strategy works.