Hydrometer Recommendation System and Predictive Analysis of Losses in Telemetry Networks

Abstract

Recommendation systems are essential tools that can assist in evaluating telemetric data by identifying patterns and trends within the telemetry data. In this study, we present an analysis conducted on the water network data of a municipality in the southern region of Brazil to detect patterns and develop a hydrometer recommendation system based on user profiles. We performed a correlational study between consumption characteristics and properties of measurement devices, aiming to identify the attributes with the most significant influence on water usage. Fromthe descriptive exploration of the data, we observed that both the type and model of the hydrometer and the age of the device could impact user measurement results and consumption behavior. Consequently, the analysis results were utilized to create a system capable of identifying and visually confirming the intersections between the desired characteristics.