Navigating Contextual Complexity with Graph Visualization
In recent years , the observability of information has exploded , with new sensors , algorithms , and machine learning summarizing and expanding the available data in nearly every domain . Orientation in these increasingly complex data spaces can be difficult . Maintaining orientation when navigating within those spaces is a further complication .
Intuitively , sighted persons know this as users of maps . If you don ’ t have both a clear idea of where your starting point is , as well as where your target destination is , you may have trouble orienting your mental model with the map view . As users of information systems , we also know that “ you can ’ t get there from here ” is not only a real-world wayfinding problem , but also a common user interface nightmare that can interfere with both navigation and orientation .
It is also true that we don ’ t always know where we are going when we start a journey , but we do have a starting point “ You are here ,” and even researchers generally begin with a hypothesis which sets a direction . It is important to be certain when “ you can ’ t get there from here ” is the correct answer to a specific question , and not a flaw in the affordances that have been provided to the user [ 6 ], either in the underlying graph , or in the presentation of that graph .
Graph-based data management and graph-based visualization offer new tools that afford users new options for orientation and navigation across large data systems . They also offer system designers new ways to confound users with unnavigable information spaces . If we fail to provide tools to manage complexity effectively , the user may be left navigating in a dense , dark forest . If we are too heavy-handed when managing complexity , we might hide valuable information the user needed .
1.2 COMPLEXITY FACTORS
Filipov , et . al ., summarized a set of network visualization disciplines in their meta-survey and roadmap-based discussion of the state of the art of network visualization research . [ 6 ] Their work identifies the specific special considerations for graphs that are very large or multi-faceted , with special techniques for multi-facet needs that include time , geospatial aspects , or layering of concerns . These are all common requirements of graphs that represent systems .
Statistical Structural Layered Spatial Temporal
Emphasizes clusters of connected or similar nodes , based on characteristics of the population , which may be calculated .
Emphasizes the shape or cyclic nature of connection patterns in the data .
Isolates abstraction layers through filtering , expanding / collapsing , swim lanes , or other
Superimposes a graph on geographical or other domain specific spatial maps ( e . g . buildings , processing plants ).
Synchronizes with timelines or otherwise provides a temporal map of change over time .
Table 1-1 : Data abstractions that may require specialized visualization to aid human comprehension .
72 February 2025