An alluvial plot generator with gradients

This visualizer is a basic alluvial plot generator, illustrating the use of gradients in category-based visualizations with a hopefully useful working example.

See this article for a quick definition of the profound difference between an alluvial plot (which this example creates) and a Sankey diagram (which is a whole other affair).

An alluvial plot generally tracks the distribution of a number of discrete items across a set of dimensions.  In the example data for this tool, the items are portfolios of homes to be sold, and the dimensions are properties of the homes — though you can put in whatever dimensions you want, and as many as you want.


Note that for most simple inputs, this 'optimization' doesn't do anything useful.
Drag and pinch to adjust view. Mouse over to highlight a fact, or the set of facts with a given dimension value.


Each line (other than the header) is a stripe in the alluvial plot. The header specifies what dimensions (i.e. columns) exist.

The format of each line is:
label, weight, (position in dimension 1), (position in dimension 2).... etc.

The tool on this page uses gradients (optionally) to try and visually relate the leftmost and rightmost dimension.  Given the example data, the aim is to relate ‘home type’ and ‘condition’ despite the other dimensions present — so we can get a feel at a glance for how much of our total portfolio is in condition ‘new build’, and how closely that condition correlates with, say a home type of ‘flat’.

One of the color schemes given here, ‘redlook’, is part of an experimental visualization style that’s intended to show relationships that may cross each other, without using transparency or blending.  It’s not really practical but I love the starkness so I’ve left it in.

Tagged : / / /

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Articles