Visual Analytics: Definition, Process, and Challenges - Information Visualization: Human-Centered Issues and Perspectives - page 170

Fig. 7. The temporal histograms show the distribution of the stops in the frequently
visited places (Figure 6) with respect to the weekly (left) and daily (right) cycles.
a map (Figure 6) or 3D spatial view. Temporal histograms (Figure 7) are used
to explore the temporal distribution of the stops throughout the time period and
within various temporal cycles (daily, weekly, etc.). These complementary views
allow a human analyst to understand the meanings or roles of the frequently
visited places.
In order to detect and interpret typical routes of the movement between the
significant places, the analyst first applies a database query to extract sequences
of position records between the stops, from which trajectories (time-referenced
lines) are constructed. Then, clustering is applied with the use of specially de-
vised similarity measures. The results are computationally generalized and sum-
marized and displayed in the spatial context (Figure 8).
7.2 Multilevel Visualization of the Worldwide Air
Transportation Network
The air transportation network has now become more dense and more complex
at all geographical levels. Its dynamic no more rests on simple territorial logics.
The challenge is to gain insightful understandings on how the routes carrying the
densest trac organize themselves and impact the organization of the network
into sub-communities at lower levels. At the same time, subnetworks grow on
their own logic, involving tourism, economy or territorial control, and influence
or fight against each other. Because of the network size and complexity, its study
can no more rely on traditional world map and requires novel visualization. A
careful analysis of the network structural properties, requiring recent results on
small world phenomenon, reveals its multilevel community structure.
The original network is organized into a top level network of communi-
ties (Figure 9(a)). Each component can then be further decomposed into sub-
communities. Capitals such as New York, Chicago, Paris or London (Figure 9(b))
clearly attract most of the international trac and impose routes to fly the world
around because of airline partnerships (economical logic). Asia (Figure 9(c))
clearly stands apart from these core hubs because of strong territorial ties en-
dorsed by national Asian airline companies (territorial logic). Visualization of
social networks such as the worldwide air transportation is challenged by the
necessity to scale with the growing size of network data while being able to offer