This page is from our blog archive
It’s still useful, but check out automatic layouts for more up-to-date info.
A couple of blog posts ago, we introduced our network clustering algorithm – to help uncover community structures in your data. At the time, we skipped over the cool new layout featured in the screengrab. Let’s take a closer look at…
The Lens Layout
Lens layout is another example of a force-directed layout, meaning it aims to produce a network layout with consistent link lengths and minimal overlaps. Specifically, the lens layout uses a combination of the repulsion/spring system in our standard layout, and an isometric approach to uniformly spread nodes within the circular drawing space.
The result is a layout algorithm that is computationally efficient and allows us to get a good ‘close up’ view of the nodes:
Data taken from this research paper.
How does the lens layout work?
The Lens layout is inspired by the Binary Stress (bStress) model, created by Yehuda Koren and Ali Civril. We have made a number of modifications and optimizations, but the overall approach is similar:
- The nodes are uniformly arranged into a circular grid
- The algorithm clusters connected nodes in varying degrees of tightness
- The most efficient equilibrium between clustering and uniform spread is reached
As a consequence of this, we see that places densely connected nodes move towards the center, while less densely connected nodes are pushed to the periphery.
Why should I use it?
There are two scenarios in which this layout is particularly useful:
Large networks
The lens layout is relatively computationally cheap. That means it can be more easily scaled up to very large networks.
Disconnected networks or networks with singletons
Singletons and disconnected networks are treated the same way as connected networks. That means no packing algorithm is needed to retrieve and re-position far-flung nodes.
Find the right layout for the job
Analyzing and visualizing networks can be a case of trial and error. The more views you can have of your network topologies and dynamics, the more likely you are to understand what’s happening.
Why not try our graph visualization products on your own data? Request a free trial to get started.
