WRc’s specialist data science team work closely with engineers, scientists and social scientists to gain greater insights using existing data sources such as asset information, weather, population and community information. Increasingly new machine-learning techniques can be applied to leakage, night use and customer demand investigations to improve understanding of factors that influence these. This allows models to be developed that have good predictive power for a variety of long term planning, and short term operational decision making.
Our team has developed cutting edge leakage prediction models that do not rely on preceding leakage, and hence sister models based on the same underlying data can be developed, for instance to reliably forecast the type, as well as scale, of leaks that a given DMA is sensitive to under different weather conditions by changing the output variable to bursts. Or, by integrating social information on a wide range of issues from typical employment patterns to school holidays, allowing separation between leakage response and night use response so that our customers are in the best position to make strategic decisions regarding leakage or customer communication interventions.
Get in touch with our team to explore them any applications of data science to ensure the most value is achieved from your investment in data collection.