This work presents ArsEmotica, an application software that aims at extracting a rich emotional semantics (i.e. not limited to a positive or a negative reception) of tagged resources. We worked on a subset of artworks supplied by the art portal ArsMeteo, and focussed the sentiment analysis on the collections of tags related to artworks. Our approach exploits and combines multilingual lexicons (MultiWordNet), affective, and sentiment lexicons (WordNet-Affect, Senti-WordNet) with an ontology of emotional categories (OntoEmotion), enriched with over four hundred Italian emotional words referring to the about eighty-five concepts of the ontology. ArsEmotica uses the ontology to identify tags directly referring to emotions, while potentially affective tags can be emotionally annotated by the user, in a pure Web 2.0 approach. From all these different information sources, ArsEmotica ranks the emotions related to the artworks.