Ontology-Driven Visual Analytics Software Development

This paper is devoted to the automation of ontology-driven development of visual analytics applications. Ontology-driven tools for constructing data flow diagrams (DFDs) are proposed; these tools are used to describe data analytics pipelines in the SciVi platform. Both the development environment and the generated visual analytics applications are ontology-driven. A unified application ontology model for automated code generation taking into account the specifics of target programming languages is proposed. The extensible SciVi repository of ontologies is based on this model. When the repository includes ontologies adequate to a specific application domain and the DFD model is suitable for a desired data processing pipeline, it is possible not only to automate the development of visual analytics software but also to guarantee its validity.