Antology-Graph

This dataset provides a directed graph of semantic proximity between intellectual concepts (philosophers, scientists, artists, educational institutions, religions, ideologies, and fields of knowledge), built from the mappingbased-objects dump (version lang=en) of DBpedia. The graph was produced as part of the project for the course MC859 - Project in Theoretical Computer Science (University of Campinas, 2025), which applies centrality algorithms (PageRank and Personalized PageRank) and community detection methods to open ontologies.
Starting from the 22,791,171 semantic triples in the original dump, a subnetwork was extracted based on 11 thematic predicates selected for their relevance to the analysis of intellectual prestige flow: dbo:influencedBy, dbo:influenced, dbo:doctoralAdvisor, dbo:doctoralStudent, dbo:academicAdvisor, dbo:almaMater, dbo:knownFor, dbo:notableWork, dbo:field, dbo:religion, and dbo:ideology. Inverse relations (influencedBy, doctoralAdvisor, academicAdvisor) were semantically reoriented so that all edges point from the influencer to the influenced entity, aligning the graph with the canonical interpretation of PageRank as a measure of prestige. After reorientation and deduplication, the final graph contains 326,270 vertices and 426,617 directed edges, with an average degree of 2.615.
The graph is highly fragmented in terms of strongly connected components (approximately 325,000 SCCs, of which 99.97% are singletons), but structurally cohesive when weakly connected components are considered — a typical characteristic of directed ontologies that encode chronological relations of influence. The degree distribution follows a power law (scale-free network), with prominent hubs at educational institutions (Harvard, Cambridge, Yale), religions (Christianity, Islam, Catholicism), and disciplines (Painting, Philosophy, Law).