Water Assessment Using Geospatial and Data Science Tools

Abstract

The main objective of this study was to determine the surface water and soil moisture available on Earth, and to test water quality using geospatial and machine learning (ML) tools. Java and Python scripts were developed to design the model. This study presents a smart approach for collecting and assessing water bodies present on Earth. In this study, we identified the surface water and soil moisture sites on Earth and subsequently identified the surface water and soil moisture sites in Taiwan. To test the quality of the water, we designed an ML model. Up on experiment, the random forest model obtained training and test accuracy scores of 100% and 68%, respectively. To improve the test accuracy score further, we used the auto-ML technique and obtained a test accuracy score of 69%. Therefore, based on the accuracy scores, we concluded that the auto-ML model was the best.