一种用于数据质量规范与操作化的模型驱动管道：面向领域专家的无代码方法
A Model-Driven Pipeline for Data Quality Specification and Operationalization: A No-Code Approach for Domain Experts
浏览论文内容
中文总结 AI 辅助
针对领域专家缺乏技术专长难以将数据质量期望形式化的问题，提出基于QPM的模型驱动管道，通过Constrainify网络应用定制模板并转化为可执行分析，实现数据质量约束的形式化与操作化，得到可重复使用的特定领域质量分析。
中文摘要 AI 辅助
高质量数据对跨领域可靠分析、决策和研究至关重要，在文化遗产等领域尤其如此，人工收集整理的数据易出现质量问题。为提高数据质量需定期进行系统质量分析，由领域专家用自然语言表达期望，但他们缺乏将其形式化的技术专长，导致该过程耗时且对技术要求高。为此提出一个形式化和操作化数据质量约束的管道，用QPM支持，通过Constrainify网络应用定制模板并转化为可执行分析，得到可重复使用的特定领域质量分析。
英文摘要
High-quality data is essential for reliable analysis, decision-making, and research across domains. This is especially relevant in areas such as cultural heritage, where data is collected and curated manually, making it prone to quality issues like inconsistencies. To improve data quality, the data must be analyzed regularly using systematic quality analyses. Quality analyses validate the conformance of data to domain-specific expectations. These expectations are best understood by domain experts, who can express them using natural language. However, they rarely possess the technical expertise to formalize these expectations into executable quality analyses. Consequently, this process requires domain experts and data engineers, making it time-consuming and technically demanding. The required technical expertise and the resulting dependencies pose a significant challenge. To address this challenge, we present a pipeline for formalizing and operationalizing data quality constraints. We support this pipeline using QPM, a metamodel for defining templates for reusable quality analyses. The web application Constrainify enables tailoring templates to specific conceptual requirements and translating them into executable quality analyses via a tool-chain based on model-driven engineering subpipelines. The result is a set of reusable, repeatable, and domain-specific quality analyses.
