Navigating Contextual Complexity with Graph Visualization
Figure
4-2 : Hierarchical visualization of a sample RDF model of IFC4x3 .
The IFC ontology is extensive , with more than 1500 defined classes and more than 7000 edges . The hierarchical visualization of this ontology tends to be quite flat , which makes it challenging to identify patterns , or to navigate effectively , as seen in Figure 4-2 .
By contrast , the circular layout in Figure 4-3 highlights the distinct modeling patterns at work in this ontology and more readily affords an investigation of sections of the model that indicate unusual patterns , or lack of patterns . Approaches for automatic graph layout , such as the results we see in this circular layout , address several NP-Complete and NP-Hard problems such as minimizing the number of edge crossings , minimizing total edge length , and minimizing the number of inter-cluster edges . Due to the nature of the graph layout problem , automatic graph layout technologies use heuristic approaches to produce quality visualizations . See [ 16 ] for further discussion .
Figure 4-3 : Circular layout of the IFC4x3 sample ontology showing distinct modeling patterns .
The labels in Figure 4 – 3 show patterns related to these portions of the model :
• Elements near label A represent drawing primitives and spatial elements
• The large group near B is the set of primitive measures
• Groups near C represent domain-specific portions of the model .
• Disconnected nodes near D are resource level elements .
• The group of disconnected nodes at E are Enumerations .
This illustrates the distinct separation of the drawing layer ( A ) from the domain concept layer ( C ). While views such as these improve the chance that interesting patterns , or deviations from
80 February 2025