Trees & Graphs

Trees are nothing more than restricted types of graphs, just with many more rules to follow. A tree will always be a graph, but not all graphs will be trees.

Trees always have a root node and follow a simple parent-child relationship.

A tree diagram or graph is best used for a sequence of events - it's hierarchical and shows the relationships between whatever value is inputted. It's often used with fractions to show probability.

'Position' is a key visual variable in a tree diagram, less so in a graph, as every value's position relative to another is what represents the hierarchy. In a graph, values can basically be anywhere so long as they're visually connected in another way.

Other visual variables can be used to show classifications / groupings within the overall structure, how connected values otherwise differ.

As far as pre-processing, there doesn't seem to be any specific methods needed for trees & graphs so long as the data can be appropriately analyzed (if labels instead of purely numerical) and the data is categorized.

Graphs are networks, where multiple nodes can connect, but they can be directed or undirected.

'Good' Examples

Giorgia Lupi 'Bruises', conveys something unconventional and does it clearly and beautifully

This example uses color to convey additional information than just hierarchy

Employs color, typography, and a visually interesting layout

'Bad' Examples

Using a different visualization type *inside* a tree is perhaps not a great idea (too confusing)

Differing line weight is a good idea, but confusing when only used for one-half of variables (male, not female)

This is visually interesting but too much information to learn anything about the individual values or relationships