Layout

The 2D graph layout can use a selection of layout algorithms to achieve different visual results, in some cases having very different running times.

cola

Produces the most attractive layout for most graphs.

The CoLa layout algorithm is implemented with WebCola, developed by Tim Dwyer of the Monash Adaptive Visualisation Lab (MArVL).

cola-flow

A variation on the CoLa layout for directed graphs, which will arrange nodes so that edges run generally downward.

cose

The COSE (Compound Spring Embedder) layout algortithm is a physics-based method that produces fast results for large, complex graphs.

cose-bilkent

A variation on the cose layout developed by the i-Vis group of Bilkent University. It aims for a good balance between looks and speed

grid

Simple gridwise positioning with no meaningful edge handling. This will provide the fastest possible result, but will have limited utility for complex graphs.

concentric

Simple positioning into concentric circles. It has similar performance and drawbacks to the grid layout.

Bundle

The bundling option is handled differently by each layout option, but generally attempts to group nodes of the same value for the given attribute.

Controls

Use shift + click or shift + drag to select multiple nodes, which can then be moved together.

Nodes that have been bundled together can be moved as a group by hovering with their area until a box around the group is visible, then dragging.

This visualisation places nodes around a circular arrangement, with edges crossing the circle and optionally bundled by value.

It is an adaptation and extension of the D3.js example for hierarchical edge bundling.

Options

Display all nodes

Toggle whether to include or exclude nodes with no connected edges under the current filter and edge visibility settings.

Bundle

Group nodes together based on the given attribute value. Edges wil also be bundled together to better differentiate connectivity between bundle groups.

For continuous values this will group into bands.

For attributes where a node can have multiple values (e.g., the assignment of a single node to multiple modules in an overlapping module decomposition), nodes will be grouped with other nodes that have exactly the same set of values. Nodes with more values (e.g., belong to more modules) will be placed in concentric circles closer to the centre.

Sort

Assign an attribute used to determine the order of nodes around the circle. If bundling is applied, this will affect the order within a bundle, but not the bundles themselves.

Label

Set the node label to the label data, or alternatively the value of a given attribute.

Histogram

The histogram option allows for the addition of a bar graph representation of node attributes to be added on the circle's outer edge. Set the number and assignment of attributes, and use the color palette to set the respective colours.

The model used to represent the brain surface can changed to a preset model or from a custom model file uploaded by the user.

To use a custom model, select the "Upload" option, specify the file to use with the "Choose File" button, and finally press "Upload" to transfer and apply the new model.

The model must be in Wavefront OBJ (.obj) format. Some 3D modelling software such as 3ds Max can import and export to this format.