How are we going to prepare the water sector for data-driven decision-making?

Traditional decision-making often relies on intuition, experience and theory. In the strategic decision-making data can play an important role. Data-driven decision-making is guided more by facts, statistics and data. Data-fed models can thus offer insight into the different consequences of decisions. Data can also help in providing an idea of the uncertainties in the expected results. This enables well-founded decisions, for example, concerning the use of sources, treatment methods, transport pipelines, new production locations and operational management.

Data-driven decision-making: the background

KWR researcher Peter van Thienen opened the meeting with a description of the history of data analysis and artificial intelligence. Data analysis for decision-making purposes began early on; for example, during the cholera outbreak in 1854, John Snow made use of data analysis to identify the source of the water that was making people sick. The objective of data-driven decision-making is ultimately to analyse data automatically, thus providing one with the insights with which to make decisions. This involves decision-making at three levels:

• Strategic: relating to the system choice
• Tactical: system design
• Operational: system performance.

Different sources of uncertainties play a role here. The uncertainty is limited at an operational level, but it is greater at a tactical and strategic decision-making level. One can therefore easily automate operational decisions, but human interpretation is more important in strategic and tactical ones.

Figure 1: Characteristics of a data-driven organisation.

Dynamic adaptive pathways (Deltares)

Deltares has developed a staged approach for decision-making with the help of data. This was the subject of a presentation during the meeting by Ad Jeuken from the knowledge institute. Dynamic Adaptive Pathways is a framework for resilience and flexibility under changing conditions. It makes it possible to analyse pathway dependencies in risks for long-term modifications to infrastructure. The analysis reveals the impact of actions by means of indicators and threshold values in a system. The outcomes of different sets of actions over a time span can then be outlined. These can be weighed against the cost of the actions. Ultimately, an adaptive pathway can be laid out, which specifies which actions one needs to take and when, with a view of producing an optimal result. Jeuken explained this approach using a number of model projects. Deltares has applied this methodology in analyses conducted for Vitens.

Dike inspections (WDODelta)

Wijnand Evers from WDODelta gave a presentation about how data helps in dike inspections. WDODelta has developed the Water Safety Dashboard, which contains for instance the current water levels, precipitation data, structures, high-water concern areas, current strength and historical data. Thanks to the tool the defence managers can be sent to the right locations. The dashboard also presents information about where the dike system is still not optimal. Measurements taken by the Water Authority during dike inspections are automatically recorded in ArcGIS. On the basis of these data, reports can be sent directly to a contractor whenever maintenance work is needed on a dike. When a report is dealt with and the work completed, the information is recorded in the database history. Decision-making is facilitated thanks to this easy access to information. The provision of information is especially crucial in a crisis situation.

Energy transition and digitalisation (TNO)

The last presentation at the meeting was given by Devin Diran of TNO. He spoke about the role of digitalisation and data-driven work in the energy transition. This instance also involves the use of a dashboard to make the energy transition clear and comprehensible. Data-driven decision-making gives municipalities better insight into system-dependencies. Predictive analyses are possible, but so too is the monitoring of progress in the transition and of the effectiveness of policy. Diran drew on a number of cases to illustrate the fact that one learns by experimenting. Moreover, the work is both multidisciplinary and cross-domain in nature.

Discussion among water utilities

The meeting concluded with an interactive discussion on data-driven decision-making at drinking water utilities. The key question here is: What is the first step? To begin with, the digital transformation involves the creation of new roles in the organisation, which need to be filled by people. Management could also start working in a data-driven manner, though the task is often seen as being purely operational. How can we prepare people for data-driven decision-making? The solutions proposed focus on connecting people in the business with data scientists, the professionalisation in data processing, and the establishment of quality assurance for data-driven applications. A survey of the meeting’s participants shows that the Dutch and Flemish drinking water utilities see the importance of data-driven decision-making, but preparatory steps still need to be taken.