Navigating Contextual Complexity with Graph Visualization
• Graph designer or ontologist : A well-described system requires a well-described , and possibly domain-specific meta model or schema . A graph schema designer needs to see and manipulate all the details of the meta-structure of the graph . The meta-modeler needs to understand the finer points of logic and the effect that rules and constraints have on the behavior of the resulting instance graph ( s ). When ontologies or other schema get very large , tools that apply to the data scientist / analyst may be needed .
• Data scientist / analyst : The analyst tends to act on vast amounts of data to generate insights or to perform visual analytics , as well as use graph queries , graph analytics and advanced features of graph visualizations to arrive at insights . In many big data analysis cases , graph analytics are used to generate graph layouts that resemble and are perceived as scatterplots . While the data source may be a knowledge graph , individuals are less important than population effects and statistics . Data scientists may translate the results of these explorations into new knowledge to be acted on by other users .
• Knowledge Graph User : A knowledge graph user may not know or care that data comes from a graph-based source . They are tasked with getting something done quickly and correctly , whether that is to describe a system that complies to the meta-model or using the knowledge graph to draw a conclusion about a system . These users tend to work on smaller amounts of task-specific information , presented in a fashion designed for clarity and efficiency , and possibly in support of a specific process or domain .
Each of these users has a need for intuitive graph navigation through the system . How complex this navigation becomes is dependent upon the complexity of the information space they need to navigate , and the task they are engaged in . The following sections will describe examples that apply to these users in system engineering or similar disciplines .
• Graph Designer : Navigating and understanding large meta-models
• Data Scientist : Navigating and interpreting large graphs
• Knowledge Graph User : Navigating systems knowledge graphs and digital threads
4.1 NAVIGATING AND UNDERSTANDING SYSTEM META-MODELS
Meta-data design tasks are distinctly different than modeling instance data . This is as true for graphs as for relational data , and as true for large systems as for any other type of graph data .
Triple stores and graph databases offer new ways to store and query meta-model information , but the process of visualizing the meta-level concepts , links , and rules is distinctly different than for the users on the other end of the data pipeline , as described in section 4.3 .
Graph schema editors require the basic data engineering tools for creating and modifying the structure of the meta-model : supplying the details of a schema , adding and modifying nodes , edges and attributes , or describing the rules and constraints on the meta model . The meta-model
78 February 2025