Skip to end of metadata
Go to start of metadata

Let's look at a much larger sample graph (|V|=1695, |E|=3889), 'sample_graph/'.

Create a hive plot with these changes to the default settings:

  • use "degree" to assign node positions
  • normalize node positions

Refer to Basics - Part 1 if you don't remember how to configure the settings to create an hive plot.


Notice that the color of some nodes/edges are more saturated than others?

Their positions are just more popular than others!

Normalize opacity of nodes and edges

The normalized opacity of an item indicates the popularity of the item's position.

For example, the most opaque items are at the most popular position while the least opaque items are at the least popular position.

By default, the opacity values for both nodes and edges are normalized between 50% (the minimum opacity set by the sliders) to 100% (the maximum possible opacity).

Without normalization, opacity of stacking items would simply add up until it is maxed out at 100%. For example, it would only take 2 stacking items to max out the opacity when the minimum opacity is set to 50%.

Normalize node size

The normalized node size indicates the popularity of the node's position.

For example, largest nodes (radius of 30 pixels) are at the most popular position while smallest nodes (radius set by the slider) are at the least popular position.

By default, node sizes are not normalized.

Let's normalize the node size:

Since node opacity is normalized, the largest node is the most opaque.


When each node is given a unique position in a hive plot (ie. when positions are rank ordered or when node name is used for position assignment), there are no stacking nodes and thus normalizing node size and/or opacity would have absolutely no effect.

Fan out edges

Instead of stacking edges at the same position, edges can be fanned out.

By default, stacking edges are not fanned out.

Let's fan out the stacking edges:


Normalize edge opacity would have no effect when stacking edges are fanned out.

  • No labels