Graphics Reference
In-Depth Information
Depending on the software, sometimes nodes can be grouped together,
moved, and then ungrouped/expanded as a way to move around sets of
nodes.
Iterative Analysis
Exploratory analysis is iterative—you visually inspect, form a hypothesis,
narrow down via filtering or selection, identify, and repeat. Within this
process, changing the layout, visual attributes, and sometimes fixing bad
data are also part of the iterative process. Through the process of analysis
and iteration, you can become more familiar with the graph and the data,
which, in turn, spurs additional questions and can yield successive insights.
In the market basket analysis, various insights can result from the analysis.
For example, you could use the two clusters near the top-right corner in
Figure 6-6
to justify a standalone store—these products do not have a strong
reliance on other product categories. Or, you could see that topics connect
back to some more general products—the topics stand relatively alone, but
if they were located more closely to the general products, perhaps they
would sell more. Or, you might see that general hardware products could be
sold with the do-it-yourself topics in bookstores. With successive iterations,
insights can be improved or new insights can result.
For many of these interactions (particularly when exploring connections,
paths, and neighbors), more pixels can help make the details visible. For
example, one of the authors' workstations provides double the resolution of
current HD displays, and another workstation has five times the resolution
of HD, as shown
Figure 6-13
. Thus, the different workstations can aid in
different interactions.
