Node-Link Diagrams
Last updated
Was this helpful?
Last updated
Was this helpful?
Node-Link diagrams use nodes (points/symbols) to denote objects and links/edges (lines) to represent relationships.
Nodes are arranged according to the attraction and repulsion forces between the nodes.
The main goal is to visualize the structure of the network.
Nodes are first initialized in random positions
For each node, the sum of the forces acting on the node are calculated; this value gives the direction and intensity for moving the node
Nodes are moved accordingly
2 and 3 are repeated iteratively until a stable positioning is found
Various attributes can be encoded in nodes and links:
Color and Shape can be used to encode categorical information in the nodes
Size can be used to encode quantitative attributes in the nodes
Edge Thickness can be used to encode quantitative information in the links/edges
Patterns can be used to encode categorical information in the links/edges
Color and Arrows can be used to encode direction of the edges
In a fixed layout, the positions of the nodes remain fixed from the beginning, according to certain criteria.
There are 3 types of fixed layouts:
Circular (shown below)
Nodes are arranged in a circular fashion, and are equidistant
This is the most common fixed layout
Linear
Nodes are arranged along a line, and are equidistant
Grid
Nodes are arranged in a grid
The Force-Directed Layout, as discussed earlier, is used when the structure of the network is to be visualized. The Fixed layout is used because of the following reasons:
Visibility
Since the nodes are arranged at fixed positions, they are always visible. So are the links
This also means that the properties of the nodes and links are clearly visible
This is not the case in force-directed layouts, where some nodes may be more visible than others
A fixed layout also makes it easier to spot large and small nodes and to determine which nodes are well-connected or poorly-connected to other nodes (degree)
Grouping into Categories
A fixed layout allows the grouping of nodes into meaningful/useful categories
This allows us to visualize not just the connections between nodes, but also the relationships between different categories
When we have too many edges that cross over each other and create clutter, we use the concept of edge bundling. This refers to the bundling together of edges that flow in a similar direction, with the hope of reducing clutter and being able to clearly represent the properties of the edges.
Instead of connecting nodes using a straight line, the line follows a path that is designed by some structure that connects the nodes together. This is denoted using a spline curve.
This is used when the fixed layout is actually a map, or some other representation of space. The nodes represent fixed spatial locations.
A typical example is a migration map.
The points used to create the spline are called control points.
