Financial Cartography: Network Layouts

Kimmo Soramäki and Samantha Cook

The placement of nodes in a network visualisation can greatly affect how informative the visualisation is. For example, the placement of nodes in Figure 6.3 emphasises the importance of the Canadian dollar, and the other nodes are arranged in a circle sorted by their correlation to CAD, making the strength of relationships easy to see and compare. One strategy utilised in various layout algorithms to best show network structure is to place the nodes so as to minimise the number of link crossings. Depending on the type of network and the goals of the visualisation, however, other properties may be more desirable. For example, in a network with a strong community structure the most important factor may be closeness of nodes in the same community. This chapter discusses many different layout algorithms for network visualisation, with real data examples of each.

We will illustrate the layout algorithms presented in this chapter using a network of annual remittance data (focusing on the year 2012) compiled by the World Bank.¹¹ See http://www.worldbank.org/prospects/migrationandremittances. The links in the network represent net remittances from one country to another. For example, if