S Zhong, N Turich, P Hayde
Publication Date (Web): 15 February 2016

SA Water collects and stores large amounts of customer meter data on a proprietary customer-billing platform. To inform infrastructure planning, SA Water Corporation has a need to generate analytics based on this data. Historically, the problem with meter data was the volume of data and general accessibility. These obstacles made data processing and analysis both labour-intensive and time-consuming.

In the current regulatory environment, this approach was no longer acceptable, and a more timely and cost-effective approach was required that could provide whole-of-system outcomes. This paper provides details of the automated process developed by SA Water Corporation to analyse this data both temporally and spatially and for publishing the analytics.

Highlights include the following:

  • 20 years of billed consumption data across the entire state was analysed.
  • Analysis included drawing correlations with external economic data.
  • Detailed analytics were created and projections generated across all spatial scales.
  • Analytics allows ranking and prioritisation of need for investment based on demand.
  • A new automated process was developed using ArcGIS and Microsoft platforms. The automated analysis can be broken down into four major steps:
    Base Data and Spatial Join, Data Extraction, Consumption Analytics and Report Publication. 

The process draws on both internal metered billing data and external data including land valuation and Australian Bureau of Statistics Census data that is available in the public domain. The new process creates the ability to carry out annual detailed analysis and publication of information for metro water treatment plant areas, metro suburbs, regional districts, regional towns and groundwater supply areas, supplemented with further socio-economic analysis of the metro suburbs.

This has proven to be a significant time-saving advancement for what would previously have been a time-consuming task. The code used for publishing reports can be easily adapted to include future change in analysis.

Based on the results, SA Water is now able to rank all areas by a number of different criteria, draw correlations with external socio-economic data, visualise the spatial disparity in growth, compare the relative importance of major customers in their respective areas, analyse historical trends and automatically generate future projections.

Many insights were made from the information generated, which includes the ability to rapidly quantify long-held anecdotal information. Internal business units now have readily available data in a spatially relevant format to temporally monitor a suite of factors, highlight changes and update the decision-making processes accordingly.

In conclusion, a rapid, low-cost and robust automated analysis tool was created, based on metered water consumption data, which will contribute significantly to the business intelligence of SA Water. The analytics will allow SA Water to make better-informed infrastructure planning decisions within a regulatory environment.

The results will inform our planning resources, our understanding of available capacity, and provide insights into customer behaviour and the reprioritisation of asset investments.

Click here to read the full paper