Data visualization in three dimensional space

Two dimensional graphs are ubiquitous and thus have been researched extensively. However they still scale badly as the amount of nodes and edges increases. To reduce cluttering in this thesis an approach is presented that allows the viewer to interactively visualize interesting parts of the graph also making use of three dimensions instead of just two. This way the amount of data is reduced as well as spread out further to simplify an otherwise very complex graph, especially if it is non-planar. This approach applies to multipartite graphs and visualizes subsets that form a unary tree themselves. One example for such graphs are the contents of database tables. The user selects one subset of the graph after another to determine the unary tree to visualize. Each set of nodes is distributed inside a plane that is orthogonal to the z-axis. These planes are then again spread along the z-axis. This way all edges of the graph are distributed along the z-axis and no longer cluttered in two dimensions. Nodes of one plane are placed near the projections of their connected nodes in the plane below. This layout allows several traditional two dimensional visualization techniques to be applied to the layers as well as additional techniques that are only possible in 3D and still maintains the familiar appearance of a graph closely resembling a tree.