GSP receives a text document as input and returns an enriched document, where all mentions of places/locations are associated to the corresponding geographic coordinates. To achieve this goal, in a first step GSP performs semantic annotation with the aim of identifying relevant parts of the input text, and to link them to pertinent resources (e.g., DBpedia entities) in the Linked Data cloud. Then, GSP exploits the rich and structured information associated to RDF resources to identify, via machine learning, geographic resources and to extract the right geographic coordinates for each resource.