Detecting Quality Problems in Research Data: A Model-Driven Approach

TL;DR

This paper introduces a model-driven, technology-agnostic approach for detecting quality issues in research data, with a focus on cultural heritage datasets, using analysis patterns and tool support for XML databases.

Contribution

It presents a novel, abstract analysis pattern framework that can be adapted to various database technologies for quality problem detection in research data.

Findings

Effective in identifying quality problems in cultural heritage data

Achieved good performance and expressiveness in XML database analysis

Validated through a qualitative study with domain experts

Abstract

As scientific progress highly depends on the quality of research data, there are strict requirements for data quality coming from the scientific community. A major challenge in data quality assurance is to localise quality problems that are inherent to data. Due to the dynamic digitalisation in specific scientific fields, especially the humanities, different database technologies and data formats may be used in rather short terms to gain experiences. We present a model-driven approach to analyse the quality of research data. It allows abstracting from the underlying database technology. Based on the observation that many quality problems show anti-patterns, a data engineer formulates analysis patterns that are generic concerning the database format and technology. A domain expert chooses a pattern that has been adapted to a specific database technology and concretises it for a domain-specific database format. The resulting concrete patterns are used by data analysts to locate quality problems in their databases. As proof of concept, we implemented tool support that realises this approach for XML databases. We evaluated our approach concerning expressiveness and performance in the domain of cultural heritage based on a qualitative study on quality problems occurring in cultural heritage data.