Organising research data within complex data models is becoming increasingly important, particularly when such data is to be analysed using quantitative methods. Ontologies, knowledge graphs and semantic technologies make it possible to represent information precisely and link it in a variety of ways. Yet it is precisely this strength that presents challenges: for many potential users, such data is difficult to access. Anyone wishing to work with them often needs knowledge of the underlying ontology as well as query languages such as SPARQL.
This raises a key question: how can complex structured data be prepared in such a way that it is also easily usable for analysis or manual processing? In this blogpost, I use my own workflow to demonstrate how a graph-based data model can be transformed into a tabular, machine as well as human readable and editable dataset. The process involves not only technical transformation, but also a series of decisions required to turn a complex data model into a reliable foundation for analysis, reporting, or manual maintenance. At the same time, I will address the challenges that arose during this transformation process and present possible solutions for reconciling the various requirements of data modelling, analysis and reusability. Continue reading “Graph it, flatten it, use it: the transformation of a graph-based data model into a tabular dataset”