1.4 Ontology visualization

layers: (a) backends for ontology storage, (b) middleware for the services offered by the tool suite, and (c) client applications that access the ontology middleware and provide end-user applications. The KAON tool suite is completely implemented in Java and can be extended with plug-ins that provide to developer a flexible tool. The ontology editor is called OIModeler and is composed by a graphical user interface for managing the ontology.

1.4 Ontology visualization

When an ontology is developed, its visualization is very useful and helps working with the ontology itself. Different visualizations for ontologies have been presented in the last couple of years. While some of them are implemented as standalone applications, most visualizations are provided as plugins for ontology editors like Protégé. With reference to (Lohmann, Negru, Haag, & Ertl, 2014) in this paragraph an overview of the main approaches for ontology visualization is provided. A lot of approaches use graphs as the method to visualize ontologies since they seem to be the natural way to represent the structure of the concepts and relationships in a domain of knowledge. The graphs are, very often, rendered in force-directed or sometimes in hierarchical layouts, resulting in a very comprehensive visualization. However, only few visualizations methods show complete ontologies, while most focus just on certain aspects. OWLViz, OntoTrack and KCViz depict only the class hierarchy of ontologies. Likewise, GLOW represents the class hierarchy but uses a radial tree layout and hierarchical edge bundles to display additional property relations. There are few applications able to provide the developer with a more comprehensive graph visualizations that represent all key elements of ontologies and not only the class hierarchy and their relationships. For instance, TGViz and NavigOWL use very simple graph visualizations where all nodes and links look the same except for their color. This is different in GrOWL and SOVA, which define more elaborated notations using different symbols, colors, and node shapes. In addition, some applications have been presented able to depict a 3D graph visualization of the ontologies. In this regard, we can cite OntoSphere, or tools that use hyperbolic trees to visualize ontologies, such as OntoRama and Ontobroker. Differently from all those applications that use common node-link diagrams to represent ontologies; there are various applications that apply other diagram typologies to visualize ontologies.

For example, Jambalaya and OWLVisMod use tree maps to depict the class hierarchy of ontologies. Jambalaya also provides a nested graph visualization called SHriMP that allows to split up the class hierarchy into different views. CropCircles is a related visualization technique that visualizes the class hierarchy of ontologies with the goal to support the identification of “undermodeled” ontology parts. All these approaches visualize once again mainly the class hierarchy, without considering other property relations. Cluster Maps use a visualization technique that is based on nested circles and has also been successfully applied to ontologies. Instead of showing the class hierarchy, it visualizes individuals grouped by the classes they are instances of. Each individual is represented by a small circle that is shown inside a larger circle representing the class. Similar techniques are used in VisCover and OOBIAN Insight that additionally provide several interactive filtering capabilities. While offering appealing visualizations, these tools show only a selection of classes along with their instances but do not provide complete visualizations of ontologies. A powerful type of diagram related to OWL and often used to visualize ontologies is the class diagram of the Unified Modeling Language (UML). A common issue related to this kind of visualization is the need for the understanding of UML class diagrams which is generally familiar to most users with an IT background but not for those coming from other domains, who could find difficult to interpret UML diagrams correctly. Since any OWL ontology can be represented as RDF graph, it can also be visualized using the common RDF notation. In this context, a very powerful visualization tool is VOWL which provides an intuitive visualization that is also understandable to users less familiar with ontologies. It uses a graph visualization and is also available as a plug-in for Protégé. A detailed description of this tool is later provided since it has been used to visualize the ontologies developed and presented in this monography. Having a look to the previous content, they can be synthetized the following characteristics related to the visualization techniques:

• many visualization applications utilize diagram for ontology visualization (graph visualization, treemap, UML), they are in 2D or 3D and focus on specific aspects of ontologies such as the ontology three. • other works implement a stepwise approach of ontology exploration, where only a root class is shown at the beginning and the user has to navigate through the visualization.