Data quality is the motor of data-driven projects

The hydroinformatics knowledge exchange meetings are organised within the Joint Research Programme of KWR and the water utilities. One of their key objectives is the sharing of practical experiences with reference to a relevant theme. The central question addressed in the last meeting was how to achieve good data quality and sufficient data availability.

Large-scale approach needed

At the start of a project, it is sometimes not precisely clear whether the data quality is sufficient and, if not, what effort is needed to achieve sufficient quality. Moreover, it is at times not clear where responsibility for data quality lies. Instead of improving individual datasets, or to making a data selection in connection with a specific question, what is needed is a large-scale, structured solution. Data have to be organised in a broad fashion, and the right approach is needed in the data generation. During the knowledge exchange meeting, practical examples were used to illustrate how these challenges can be tackled.

Data steward as connector

Over the last three years Dunea has worked on setting up a central organisation for data management.  Arjan de Bruin and Nanon Klok, data experts at this South-Holland water utility, explained the project in their presentation entitled ‘Data steward as connector’. One of Dunea’s central objectives is to give shape to data-driven processes, addressing the key challenges of digitalization, data quality and accurate data management.

In one of the first steps in the process, Dunea carried out a ‘maturity assessment’ in each of its departments. The goal was to assess the status of digitalization and all associated matters. This was followed by a variety of actions, which included setting up data models and processes for data sharing, developing a data-quality dashboard, establishing a digitalization agenda, and building a data platform. The ‘data steward’ plays an important role in all of these activities. The data steward is the link between the data owner, the work floor and the information service department. In Dunea 17 employees fulfil this special role as data stewards, each of whom supervises an own data domain. Key Performance Indicators (KPIs) were created for the improvement of data quality, the capability of data sharing and mutual contact. This motivates the data stewards to achieve the objectives: an approach that works superbly.

Data-driven working

Brabant Water is also ‘on the road’ to data-driven working, said Hans van Drunen, data architect at the water utility. Data-driven working should make it possible to take decisions with greater efficiency, speed and accuracy. To reach these objectives, Brabant Water has developed a five-year plan in which data will become a component of the company architecture. The plan’s pillars are: data, behaviour and structure. Fifteen guiding principles have also been formulated, including ‘data are of good quality at the source, and there is only one master copy’, and ‘Brabant Water strives towards standardisation’. Compliance with the legal and regulatory framework is a precondition for all these developments.

Like Dunea, Brabant Water also works with data stewards who are assigned an own domain and/or sector within the water utility. This involves applying a DAMA model for good data management (see figure), with governance as the central principle. The data quality is monitored by source (‘distribution mains’ for example). Brabant Water uses a data maturity model to assess whether the developments are moving in the right direction with regard to the level of data management. For effective data use a ‘data lake’ is being built, which is intended for the storage of good quality data and as a source for reporting. With all these steps, Brabant Water is on the right path.

Inspiring pitches

Besides the speakers from Dunea and Brabant Water, other participating drinking water utilities were able to share their experiences in data quality through short pitches. These revealed that the subject is receiving attention in all cases, and this includes how the response should be organised. The manner of implementation varies among the utilities, ranging from a governance company-wide plan to the implementation of general guidelines. It was inspiring to share experiences on this subject.

Need for sharing

As is customary, the knowledge exchange meeting closed with a plenary discussion. It emerged that there is a need for sharing when it comes to the different elements related to data quality. This for instance refers to the sharing of data models and data validation rules. But also the organisation of meetings between data stewards and data architects of the different utilities. This sharing has a stimulating impact and promotes coordination, which brings the point on the horizon – namely, high-quality data for successful data-driven projects – a little closer.

DAMA structure for good data management.