This visualizer is intended mainly for data modelling and BI tools. These tools frequently support ‘dimensional modelling’ in which a fact (such as revenue) can be sliced and diced by various dimensions such as ‘location’, ‘product’ and ‘date’. Those dimensions can themselves be hierarchies, of course…
There’s no one standard diagram to represent such dimensions, so here I present what I’m calling a ‘quill diagram’ — the ‘fact’ goes at one corner and dimension hierarchies fan out from it. Dimensions are colored so as to give them visual coherence, and a bit of background shading is used to try and tie the elements of each dimension together.
This example is a slightly simplified version of the display code I used in this article.
It’s not trivial to get a ‘quill diagram’ right and the implementation here is quite naive — it’s not hard to get two nodes to appear on top of each other! Still, I find it very useful for quickly representing a simple set of dimensions.
When multiple facts, or dimensions that can share attributes come into play, the quill diagram stops being quite so useful… but thanks to Big Data we live in an age of denormalized, flattened data models, so that happens less and less…!