Rankingdom is a Wikidata entity analysis tool for knowledge workers. Its goal is to position it within its domain by comparing it with other similar entities using several indicators. For this purpose, we propose a cooperative on-demand analysis architecture that decentralizes the computation of analyses, but centralizes their results for frugality. This architecture is based on a client-side SPARQL parallelization engine that deals with the fair use policy. We show the effectiveness of our approach with a benchmark and the deployment of Rankingdom.