Optimizing water distribution in Harare, Zimbabwe using IoT and cloud computing

Abstract

Rapid urbanization in Harare, Zimbabwe, has intensified inefficiencies in water distribution, resulting in high non-revenue water (NRW) and inequitable supply. This paper presents a novel data-driven framework that integrates internet of things (IoT) sensors, machine learning (ML), and cloud computing to optimize urban water distribution. Historical and real-time data including water flow, pressure, and consumption are collected via IoT sensors and analyzed using a random forest model for accurate demand forecasting and anomaly detection, such as leaks. The model is deployed on a secure cloud-based ASP.NET platform, enabling real-time monitoring and automated valve control through ultrasonic sensors over Wi-Fi. Evaluation demonstrates superior performance with R²=0.89 for demand forecasting and anomaly detection metrics of 94% accuracy, 91% precision, 92% recall, and 91% F1-score, outperforming baseline methods. This integrated system reduces water loss, improves supply equity, and provides a scalable and cost-effective approach for smart water management in resource-constrained urban settings. The framework offers practical insights for policymakers and utilities seeking to implement sustainable, technology-driven water management solutions in developing cities.