First Principles Selection of Social Media Visualizations

Current visualizations of social networks, such as node-link diagrams, can be ineffective because they often do not clearly demonstrate key patterns or anomalies. Meanwhile, new advances in social network visualization continue to emerge and we would like to employ these tools when applicable - by leveraging cognitive first principles, we can better determine applicability. As a solution, we propose a tool called MetaDesigner that will combine the power of existing, off-the-shelf tools for generating certain types of visualizations with a novel component that will both prepare the data for visualization and judge and rank, based on cognitive first principles, the candidate visualizations produced by existing tools, so that a good visualization choice can be made or configured. MetaDesigner will thus act as a workbench that will leverage the rich functionality already available from existing tools, but it will add complementary features that allow those tools to be used behind the scenes, automatically. MetaDesigner does not preclude the development of new visualization generation tools that can be subsequently integrated; as part of its development, we intend to provide an flexible API so that a given tool can be automatically managed by MetaDesigner.