Sankey chart

Sankey chart

Sankey charts show the flow of certain quantities from one set of dimensions to another, where the width of each line connecting two dimensions is proportional to the magnitude of the flow.

A Sankey chart consists of the following elements:

  • Node – A colored rectangular fragment representing a dimension. The values of the first dimension are placed on the left (source nodes). The nodes on the right are destination nodes.
  • Flow – A line connecting a source dimension with a destination dimension. The line thickness is proportional to the flow quantity based on the selected measure.

When to use

Use a Sankey chart to show a many-to-many mapping between two domains. For example, you can visualize the transfer of resources, materials, costs, cash flows, or web traffic. By using this chart, you can draw the reader’s attention to the largest flows, the largest consumer, the main losses, and so on.

Sankey charts are visually appealing and popular but can become unreadable if you have too many dimensions with many connections. Avoid over-cluttering and provide explanations where needed.

Data requirements

To build this chart, define the required and optional data fields as follows:

  • Dimensions – Two dimensions (the order of dimensions matters: the first dimension is a source node and the second one is a destination node)
  • Measure – One measure (to represent the thickness of the line connecting two dimensions)
  • Tooltip – (Optional) One or more measures

Use case

The following chart shows the flow of the marketing costs in each department, by year. The nodes are the years and departments, where the line thickness represents the marketing cost. When pointing to the line, you can see the tooltip showing the flow order, year ® department, and the respective measure for that flow. When pointing to a node, you can see the total measure for the dimension (year or department).


For details on how to customize your visualization, see Visualization settings.

For a whole list of visualizations, see the following topics: