Graphics Reference
In-Depth Information
Summary
Graphs come in many shapes and sizes, suitable for an extremely wide
variety of business problems. To choose the right approach, you must
understand the relative strengths and weaknesses of each, know your data,
and, most importantly, know what your objectives are. Graphs can be a
solution to a problem in themselves, or they can be the organizing
framework for small multiples of other types of visualization (such as line
charts, bar charts, donut charts, or radial indicators, as shown in the
preceding examples).
Diagrammatic relationships can be shown using expressive links and nodes,
expressing a model or high-level conceptual picture of a problem.
Hierarchies can be effectively visualized using trees or sunburst charts.
Distant,near,nested,andoverlappingcommunitiescanberevealedthrough
clustered layouts and characterized with symbols or labels. Or, they can
be grouped and summarized in more detail using computational clustering
approaches.
Flows are usually best expressed as Sankey diagrams, or left-to-right trees.
When flows are traded between each node, a chord diagram is an
appropriate choice. In spatial networks, often a schematic approach to
locating nodes and routes will help to clarify the graph. When visualizing
route statistics for a large graph, a brute-force overplotting of links can still
beeffectiveinsomecases,buttile-basedaggregationtechniquescanprovide
greater scalability and color accuracy when expressing very dense areas of a
graph.
This chapter served as an introduction of the many approaches to
visualizing and analyzing graphs for business problems. The chapters in
Part 2 describe the processes and tools used to do so, starting at the
beginning with data, as discussed in Chapter 3.
