Create a Sankey Diagram easily

If a picture paints a thousand words, Sankey diagrams does more.

What does a Sankey diagram show?

Sankey diagrams show the flow of resources such as cost, materials, population migration, from one set of values to another. It presents the sources linking them to the destinations where these resource, materials or costs are used. The items being connected are called nodes while the connections are called links.

The width of the links is proportional to the quantity it represents. Keep this in mind when reading and interpreting Sankey diagram.

Origin of Sankey Diagram?
Why should you use a Sankey diagram?
How I came to know Sankey diagram
Websites to create Sankey diagrams

Origin of Sankey Diagram?

Sankey diagram is named after one of the two authors that wrote a report on “The thermal efficiency of stream engines” back to 1898. In this report, the charts displayed the flow of steam in an actual and idealized steam plant, respectively. The figure below is the classic figure used in every origin story of these charts and is now known as the original Sankey diagram.

It is a standard visualization model used to represent flow of energy, material, heat in science and engineering and, as we see now, in life cycle assessments.

Why should you use a Sankey diagram?

Sankey diagrams are useful in data analysis particularly key in exploratory data analysis, for story telling, and data visualizations.

Sankey diagrams help you or your audience see the bigger picture and identify the dominant contributors or consumers. For a manufacturing process, you can see which streams contains the highest composition of your target component(s).

These diagrams are a great way to visualize complex processes like when you want to show multiple paths through a set of stages (e.g., a manufacturing process) or communicate many-to-many mappings between two groups or domains for instance, universities and majors.

Another reason to use Sankey diagram is that it helps to focus on the single aspect or resource you want to highlight. At the same time offers multiple viewing levels. You can get a high level view, see specific details or generate interactive views. Depending on the tool you use to create these diagram, you may have the option to drill down for more targeted details (Source: here).

Care should be taken when considering this tool for your task because, it may hide actionable insight rather than highlighting them if poorly developed. It could also be difficult to differentiate links when there are two or more flows with similar values on a Sankey diagram. Moreover, some diagrams can appear too complex and intimidating for your audience.

How I came to know Sankey diagram

I came to know and use Sankey diagram in my current role.

I may have seen this diagram during my doctoral research but didn’t know what it was. And since I don’t have a vivid memory of such experience, it don’t count. It does but it doesn’t, you get me! I digress!

I was tasked with the responsibility to update the previous diagram with new data and user interface. The problem was data used to make the then existing one was unavailable and the current data were stored in different forms at different locations.

However, these seeming bottlenecks were actual opportunities for me to learn about Sankey diagrams as well as test different Sankey diagram creation tools available.

The tools available to create this…?

Websites to create Sankey diagrams

These websites have simplified creation of Sankey diagram making it easy especially for people with not much coding knowledge. Most of these sites don’t require any programming knowledge but, if you do have the experience, there are a few that have that option.

What I learnt, pretty quickly, was that some of these websites were great for small data but when it came to medium to big data, it became too tedious to use. Mainly because you have to reformat your data to the format they use then copy and paste on their site. Not ideal when working with large data.

The websites that required coding or data upload made creating Sankey charts for my big data much easier. The catch here is that you may have to write the code yourself. That being said, some of the websites provide examples you can follow and adjust or modify to suit your situation.

Visit Cool Websites for a summary each websites capabilities otherwise, or do a web search whichever you are comfortable with.