Analysis and Prediction of Water Quality Data using Machine Learning Approaches and Exploratory Data Analysis

Abstract

Drinking Water Supply (DWS) is one of the most critical and sensitive systems to maintain city operations globally. In Europe, the contradiction between the fast growth of population and obsolete water supply infrastructure is even more prominent. The high standard water quality requirement not only provides convenience for people’s daily life but also challenges the risk response time in the systems. Prevalent water quality regulations are relying on periodic parameter tests. This brings the danger in bacteria broadcast within the testing process which can last for 24-48 hours. In order to cope with these problems, we propose a EDA (Exploratory Data Analysis) model for water quality assessment. This model consists of two dimensions, including water quality parameters and score. Furthermore, we applied this model to predict water quality changes in the DWS system using a Random Forest algorithm using Pycaret. For a case study, we select an industrial water supply system. The preliminary results show that this model can provide high predictions & accuracy i.e., 73.76% for water quality understanding.