Sankey Plot

A Sankey plot is a type of graph that displays subgroups structures among more than two categorical attributes, it is very much like multi-level pie chart, however unlike pie chart which display the data structure in circle, sankey display the data structure in linear flows among different attributes.

To make a sankey plot, open a new Data Viewer session, select Setup > New plot > Sankey plot . Then select any data node from the analysis pipeline which contains categorical variables you would like to view the break down of the subgroups:

You need to select two or more categorical variables to display the results. Click on the Configure gear button to choose Axes

Choose at least two categorical attributes in the selected data node to display the composition of the data structure:

The plot is composed of nodes and links:

Nodes: the vertical bars represent a subgroup. The subgroups from the same categorical attribute are aligned at the same horizonal location. The height of a node is proportional to the value the subgroup represents.

Links: the bands connecting the nodes. Width of the bands is proportional to the value from one node to another.

The specific value is displayed along with the name of the subgroup.

When click on a node, all the links connected to the node will be highlighted; when click on a link, the two nodes from the link are highlighted.

In the Axis configuration dialog, you can drag-n-drop an attribute name to change the order, the plot will update accordingly. Attributes from the top to the bottom in the dialog will corresponding the nodes from left to right in the plot.

Once you are pleased with the appearance of the plot, push Export image button to save it to the local machine or click Save button to save the Data Viewer. The resulting dialog controls the Format, Size and Resolution of the image file. The image will be saved in your favorite format (.svg, .png and .pdf).