Sankey charts: Sorting and x-axis labels

A colleague just showed me some Sankey diagrams from Domo, currently in beta. Fantastic! With Domo's killer color palette, these charts are genuinely beautiful. Flow diagrams are visually appealing as they just grab the eye and make you explore the data. I'm a fan! I've got two suggestions to offer on this beta chart type:

 

  1. It would be helpful to be able to control the order of nodes on the left and right by name or value.
  2. It would be incredibly useful to be able to label the x-axis. That's the key visual/informational difference between a Sankey and an "Alluvial" flow diagram. Here's an example I grabbed off the Wikipedia from the paper that first described Alluvial diagrams 10 years ago.

 

So much of our data is on a timeline, that being able to label the x-axis would make these charts significantly more valuable. Even just labeling the junction-points would help considerably.

 

https://en.wikipedia.org/wiki/Alluvial_diagram


NeuroAlluvial2001-2007-691x273.png

3
3 votes

· Last Updated

Comments

  • Here's what looks like a good write up of how the time series labels make all the difference. 

     

    https://towardsdatascience.com/alluvial-diagrams-783bbbbe0195

     

    In our case, we're focused very much on change and improvements over time, be that month-to-month, quarter-to-quarter, or with mile markers based on when events happened. Like we've shipped a new version, or the customer implemented a new set of procedures. The Sankey is almost ideal for this, and we've long wanted them in Domo. But, really, I can't think of a case where we want a Sankey that isn't an Alluvial chart.

     

    I appreciate that supporting this might require adding a fourth column to the data to identify the grouping pairs a row belongs to. If it's easier, Alluvial as a distinct chart type seems like a fine distinction to make.

     

    alluvial_time.png

     

This discussion has been closed.