Testing layouts

In my experience, selecting a perfect layout on the first attempt is highly unlikely; even if the best possible algorithm is chosen, it is almost a certainty that the graph display can be improved through adjusting the base settings, not to mention all of the downstream cosmetic enhancements. Knowing this makes it essential to sample multiple layouts for your network data, which can be a very efficient process for all but the largest networks.

In the next few sections, we will walk through the process of testing a few layouts using the same dataset, and will go into greater detail compared to our exercise in Chapter 2, A Network Graph Framework, by modifying settings within each algorithm to demonstrate the resulting impact on the network graph. We'll work with the Les Miserables dataset found in the Gephi samples (available when you open Gephi), as it provides an interesting network to work with, while still being small enough to easily interpret.

Our focus during this...