Due to the sheer size of large knowledge bases, users can hardly know which identifiers are available or useful for the construction of axioms or queries. As a consequence, users might not be able to express their informational need in a structured form. Yet, users often have a very precise idea of what kind of results they would like to retrieve. With SPARQR, we introduce a SPARQL learning web service which is able to learn OWL classes for given positive and negative examples.