ResearchPod Summary
In data-driven environments, ensuring data quality is critical but often hindered by a 'skill gap.' While domain experts possess the necessary knowledge to define quality requirements, they typically lack the technical expertise to write complex database queries (e.g., SQL, XQuery, SPARQL, or Cypher). This forces a reliance on IT experts, creating inefficient, resource-intensive, and error-prone workflows. The authors address this by developing a framework that allows domain experts to define quality constraints independently of the underlying database technology.
The Quality Pattern Model (QPM) is a model-driven framework designed to abstract quality analysis. It operates on the observation that most data quality issues—such as missing information, invalid links, or duplicate entries—follow recurring logical patterns. QPM allows users to define these requirements as abstract templates. These templates are then automatically translated into the specific query language required by the target database (XML, RDF, or Neo4j). By decoupling the definition of a quality rule from its technical implementation, QPM allows a single quality requirement to be reused across different database formats.
The authors evaluated QPM using research data from the cultural heritage domain, a field characterized by diverse data formats and complex descriptive requirements. Through a qualitative user study and the creation of a catalog of 73 common data quality problems, the researchers demonstrated that QPM is highly expressive—matching or exceeding the capabilities of standard database query languages. Furthermore, the study indicates that the framework's intuitive interface successfully empowers domain experts to define and manage quality analyses without needing support from IT staff.
This framework provides a scalable solution for organizations managing heterogeneous data environments. By lowering the barrier to entry for data quality management, QPM enables more frequent and accurate data validation. This is particularly valuable in domains like cultural heritage, where data is often integrated from multiple institutions, each using different storage technologies, and where domain experts are best positioned to identify quality issues but are currently excluded from the technical implementation process.
AI-generated third-party summary by ResearchPod. Not official content or an endorsement by the paper authors or affiliated organizations.