Leveraging Multi-Level Sankey Charts in GCP Looker Studio

Recently, I encountered a requirement from our leadership team to create a visualization for a high-priority use case. As I explored various options for effective storytelling, I discovered the power of “Sankey Charts” (also known as Sankey diagrams), which had been newly added to Looker Studio.

Understanding Sankey Charts

A Sankey diagram is a valuable tool when you need to illustrate the flow between two or more domains (or nodes). It visually represents the relationships between these nodes using links (or weights), which also determine the thickness of the connections.

Getting Started

Let’s dive into the process:

1. Data Preparation: Suppose you have data with different dimensions, and you want to showcase the flow between these domains. Organize your data as shown in the table below. You can store this data in BigQuery or a spreadsheet.

Current Data

2. Data Transformation: To create a Sankey diagram, we need to structure the data appropriately. Add all the States/Countries as sources for the respective countries (targets). Similarly, consider all the countries as sources for their corresponding continents. This ensures we have a two-dimensional dataset that captures the flow.

Summarise data in below format:

It should look like below

Lets not worry about weight yet.

3. Setting Up Looker Studio:

• Open Looker Studio.
• Create a blank report.
• Add your prepared data to the report.

4. Creating the Sankey Chart:

• Add a Sankey chart to your report.
• Define the source and target nodes.
• Specify the weights (if available).

5. Visualizing the Sankey Diagram: Your Sankey diagram should now illustrate the flow between different domains, as demonstrated in the example below

Important Considerations:

When working with Sankey charts, there are several crucial points to keep in mind:

1. Duplicate Rows:

• Sankey diagrams do not support duplicate rows. Therefore, it’s essential to remove any duplicate entries from your dataset before creating the chart.
• If you encounter an error message like “Cycle found in data. This chart does not support cycle in the data.” It likely indicates duplicate data. For instance, in the example provided, New York appeared as both a source and a target. Adjust your data accordingly to avoid such cycles.

2. Weight and Relationships:

• The thickness of the links in a Sankey diagram represents the strength of the relationship between nodes.
• Consider using aggregate functions to calculate weights based on your specific requirements and data context.

Remember these points as you explore Sankey charts, and happy learning!

Feel free to reach out if you have any further questions or need additional assistance!

Reference: https://support.google.com/looker-studio/answer/13412211?hl=en#sankey_charts_in_looker_studio