In every organisation, decision-making, operations, service provision and the definition of strategy stands or falls with the correct quality of the available data. Think, for example, of emergency services that cannot reach the right location on time without good data in their navigation systems, or an asset management organisation that cannot make efficient maintenance plans without good data about its assets. Wageningen Environmental Research is developing methods to check and improve the quality of spatial data in particular, from a user's perspective.
Data is increasingly the basis for decisions and work processes, so the importance of good data quality is growing. The crucial question, however, is when is data quality good enough? The quality required is strongly dependent on the purpose for which it is used. Data can be very useful for carrying out a scenario study to calculate a policy option, while the same data is not suitable for legal decisions. We call this "fitness for use". If the quality of data is insufficiently known, it is impossible to determine to what extent data is suitable to be used for a certain application.
It is of great importance to closely involve both data experts and users in the design and implementation of quality control. Wageningen Environmental Research has developed a method for this. It is important that the quality requirements are relevant to the intended use and in line with what the user considers important. Together with the user, we determine which quality requirements must be checked and what the minimum standards are for these requirements. These can be the usual quality requirements such as accuracy and completeness, but sometimes the reputation of a data supplier can also be important. If completeness is agreed as a quality requirement, it must also be determined when the data is complete enough for that specific application; is it for example 80% or 99%? Finally, we set up tests per quality requirement so that the quality checks can be carried out.
Impact and future perspective
By encouraging multiple uses of data, the assessment of 'fitness for use' will become increasingly important in determining data quality. The use of spatial data is very diverse, however, and quality assessment is therefore becoming increasingly complex. The interpretation of the concept of quality is shifting from data-oriented to context-determined. This changes it from an absolute to a relative concept. Our method starts from a different starting point, namely the user and the intended use.
It turns out that other quality requirements are sometimes more important to the user than the 'classic' quality requirements of actuality, accuracy, completeness and correctness that researchers usually use in their data quality studies. The empirical approach in our method gives users the opportunity to match quality research with their own values and ideas about quality. The method contributes to a better understanding of data quality as it is now, but also creates perspectives for improvements.