Using Murky Data to Clean up Wastewater Flooding

Annual Conference

❄

As the saying goes, “You can have data without information, but you cannot have information without data”. Many organisations are data rich; however, this data is rarely used to its full potential. This paper will describe how by pairing data with analytics, operational decisions can become more automated. This enables organisations to become better informed and make decisions backed by robust analytical data rather than solely instinct.

The paper will describe examples used in UK Water Authorities to clear up some of the “murky water” from their data - scenarios where various data sources are used in isolation for different problems. This creates a missed opportunity to gain a greater understanding of the broader issues. The approaches linked various data sources and presented them in a manner that allowed Clients to view and understand their data in a more holistic way. Enabling them to better targeted solutions to clear up the real world “murky water” problem of wastewater flooding.

In this paper we will describe how we compiled and analysed 32 years’ worth of incident data involving eight million data points across 20 different datasets for 22,000 properties. The project successfully connected and automated the analysis of the data giving the team better ways to target tried and tested solutions, along with opportunities to trial new innovative solutions.

Solutions that will be described include:

  • The use of predictive analytics, to target properties that can benefit from prioritising proactive responses;
  • Automated logical Feature Manipulation Engine (FME) analysis of historic events to recommend and quantify solutions at a regional level;
  • Collation of multiple data sources, and displaying historical data for properties with the use of ruby scripting to optimise the review of properties at an individual property level;
  • Analysis of historic events combined with a comparison of property characteristic to identify unique problems that were area specific, such as “Buchan Traps” and the implementation of specific programmes to reduce flooding in these properties.
  • Gap analysis to target CCTV and jetting in areas at risk of blockages, backed by data rather than solely reactive instincts;
  • Targeted trial of network Fat, Oil and Grease dosing within the network to prevent repeated blockages occurring.
  • Network and property analysis for installation of low-cost property level alarms, to gain an early warning of network surcharge, where previously the first signs would be from internal flooding.

    • This paper will show how even when large volumes of data are available, it is not always used to its full potential and is often lost in poor management or unusable formatting. We will present how different data analytic techniques can be used to optimally target a reduction in wastewater flooding.

    The processes utilised by this project can be applied to many other areas with operationally data rich sources to enable improved performance.

    ❄
    ❄

    2. Using Murky Data.pdf

    pdf
    1 MB
    17 Oct 2019

    1400 Parkinson_Rico_Using Murky Data.pdf

    pdf
    2 MB
    17 Oct 2019