South Australian Water:
Supplier Demand
Prediction

South Australian Water, a regional water supplier, must accurately predict daily water consumption to manage large-scale water movement across its network, a complex and expensive process. Their incumbent predictive tool was designed for oil pipeline forecasting and was poorly suited to local operating conditions. The platform lacked flexibility, was difficult to adjust, and carried high licensing costs. Our customer needed a modern, adaptable and cost-efficient solution tuned to Australian usage patterns and environmental factors.

Synechron worked with the organization’s experts to identify key drivers of water demand and engineered additional features to improve predictive accuracy. A proof of concept was developed and then productionised using Azure Machine Learning alongside Azure Blob Storage, SSIS, advanced T-SQL and SQL Server. Extensive training and documentation enabled business teams to retrain and refine models independently as patterns evolve. The new ML-driven predictions were integrated into the customer’s reporting environment, providing higher accuracy and a smooth transition from the legacy system.