Graph Visualization For Geography - Methods for Multilevel Analysis and Visualisation of Geographical Networks - page 77

Chapter 6
Graph Visualization For Geography

Antoine Lambert, Romain Bourqui, and David Auber

6.1 Introduction

Network analysis is usually performed with combinatorial measures such as those
presented in Chap. 5 or traditional statistical tools such as principal component
analysis. However, it is extremely difficult for some patterns, even with these
powerful approaches, and the results of these methods can be difficult to interpret
without an appropriate representation.
The purpose of this chapter is to present methods for the automatic drawing of
networks, which is a complex field that can raise several intractable problems related
to graph theory. However, for the purpose of geographic network analysis, there are
two main approaches that are sufficiently general for use in this application. The
first method is the so-called force-directed approach that uses a physical analogy
to construct a model network where the Euclidean distances are very similar to the
distances in the actual network. The second method creates an abstract network
that reduces the visual complexity of the representation. Because the geographic
positions of the network elements can provide crucial information, one specific
approach relates the topology of the network to the geographical positions. Then,
the flow map layout technique can be applied to emphasize high level patterns in
that network.
The remainder of this chapter is structured as follows. Section 6.2 reviews some
related work on general graph drawing methods and, in particular, force-directed
approaches. In Sect. 6.3 , we present the compound graph visualization technique.
We next explain how node positions can be taken into account and high level
patterns can be highlighted in Sect. 6.3.1 . Finally, we draw our conclusions in
Sect. 6.5 .