Rankingdom: A cooperative architecture for the on-demand analysis of Wikidata

Large knowledge graphs such as DBpedia, Wikidata, and YAGO are rich sources of structured information, widely used in domains like information retrieval and recommendation. However, their potential for supporting knowledge workers remains underexploited. As in other complex networks where specialized metrics have emerged (e.g., bibliometrics), knowledge graphs offer promising opportunities for domain-specific analysis – particularly in areas lacking established quantitative tools. In this paper, we present Rankingdom , a web application for knowledge workers to analyze any entity from Wikidata despite its heterogeneity and sheer volume. For this purpose, we introduce complementary indicators to position or evaluate any Wikidata entity within its domain, addressing analysis tasks that are challenging for traditional methods. We propose an on-demand analysis architecture that distributes computation to clients while centralizing results for frugality, by benefiting from a client-side SPARQL parallelization engine ( SParaQL ). We demonstrate the effectiveness of SParaQL through performance tests on DBpedia, Wikidata, and YAGO with two analytical queries, as well as via a real-world deployment including caching of over 10,000 entities and a user study.