Jan Polowinski: "Sharing and Reusing Visualizations for the Web of Data with the RDFS/OWL Visualization Language (RVL)"
When creating visualizations, switching between visualization tools usually requires users to set up again all visual mappings in the other tool – useful visualizations created by others are hard to reuse and combine. Furthermore, many visualization design systems do not exploit the richer semantics of data in RDFS and OWL. RVL allows visualization authors to declaratively define visual mappings from RDFS properties to graphic means. On the source side of the mappings we can select properties like “foaf:knows” or “rdfs:label”. On the target side we map to concepts from the Visualization Ontology (VISO), which offers a large palette of graphic attributes (“color”, “shape”, “height”, ...) but also relations between graphic objects such as “linking”, “containment” or “alignment”. Focusing on concise, reusable definitions of visual mappings distinguishes RVL from other approaches of visualizing RDF data that often expect SPARQL-query results in tabular shape and then map result columns to fields of fixed graphic representations (e.g. a Bar chart). The composability of visual mappings is realized by a sub-mapping mechanism, referencing existing graphic objects created by earlier mappings via their graphic role. A second difference to many general-purpose visualization design systems is that RVL is specific to RDFS and OWL. For example, sub-property and sub-class hierarchies are evaluated to find the most specific data value that is mapped. Following the convention over configuration principle, many defaults further simplify the compact definition of visualizations for RDFS/OWL data. A prototypical interpreter for a first draft of the RVL specification has been implemented and is offered open source for cooperation and discussion.
Examples can be found at: http://www-st.inf.tu-dresden.de/semvis/blog/?page_id=287
Project: https://github.com/semvis/rvl