Scientific and semantic web systems often describe the same research objects with different taxonomies, labels, and levels of detail. That makes it difficult to compare annotations, reproduce analyses, and build dependable tools on top of research metadata.
Context
My knowledge graph work focuses on scientific metadata, ontology evolution, and semantic alignment. It connects several public research threads already represented on this site: comparing annotations across scientific knowledge graphs, assessing category overlap between OpenAlex and OpenAIRE, and building ontology versioning infrastructure with DBpedia.
My role
I contribute as a researcher and engineer, combining data analysis, semantic web modeling, Python tooling, and research writing. The work includes preparing datasets, comparing schemas, writing reproducible analysis code, and explaining where automated and manual annotations diverge.
Approach
- Model research metadata as graph-shaped data rather than isolated records.
- Compare taxonomies and category systems across scientific knowledge graphs.
- Use RDF, SPARQL, and Python to inspect annotations and surface overlap or inconsistency.
- Connect ontology versioning and access reliability to reproducible semantic web infrastructure.
Related work
- Assessing the Overlap of Science Knowledge Graphs
- Are Scientific Annotations Consistently Represented across Science Knowledge Graphs?
- DBpedia Ontology Time Machine
Lessons
Knowledge graph quality is not only a modeling problem. It also depends on provenance, category design, tooling, versioning, and the ability to explain why two systems annotate the same entity differently.
Explore related work
For broader context, see Jenifer's research overview, technical skills, and contact page.