Most Charts Miss This, Alluvial Diagram Does Not

Complex Data Made Simple With Alluvial Diagram

Image By Author

in 2016, the usa used an alluvial diagram during the presidential election studies. it was part of research conducted on american voters to analyze their behavior across the country. not only that, it also shows that alluvial diagrams are useful in healthcare data analysis, business analysis, demography, and many other areas.

you can see that all these sectors deal with multi dimensional data. they have different categories, and the values often change over time.

however, some people say that alluvial diagrams are too cluttered and hard to read, even for simple cases. instead, they suggest using simpler charts like bar charts.

so today’s article is about why, when, and where to use an alluvial diagram. we will also look at when you should not use it, and what conditions you should check before choosing it.

Image By Author

for example, you can see here the 2016 usa voter behavior. there was a major shift in leader choice. more than half of obama voters moved to trump, and even many voters who did not vote in 2012 and stayed inactive also chose trump.

this simple plotting using an alluvial diagram shows how a major shift could have happened. of course, there might be many other influencing factors, but it clearly highlights that a large group of voters moved toward trump.

the next step for the researcher is to find out why this shift happened. it could be due to promises made by trump, his popularity, or what influenced republican voters to support him.

this is how an alluvial diagram helps. it does not give the final answer, but it clearly shows the flow and direction of change, which guides further analysis.

Alluvial Diagram Criticism:

we also have to look at the other side, which is criticism. it is often said that this chart is a little complex to read for a general audience. instead, a simple bar chart can be more understandable, especially in research papers.

this is partly true. an alluvial diagram becomes hard to read when the data is too large and has too many categories, sometimes even hundreds. but this problem is not only with alluvial diagrams, it can happen with almost any chart.

it is also often misinterpreted or misunderstood. when there are too many categories, it can look messy, almost like a hairball, and lose its clarity.

so now let’s look at when and how to use it properly.

Image By Author

1. look at this chart. it looks crazy and nobody wants to look at a chart like this. it is too confusing and does not give you anything clear. and for people who do not know how to read an alluvial chart, it may just look like a random design. so let’s see what you should avoid.
2. avoid an alluvial chart when the data has too many sub categories. if the hierarchy goes one after another and becomes too deep, like more than 10 levels, then it loses meaning.
3. avoid it when your goal is to compare data across decades or long time behavior. it is not the right chart for that kind of comparison.
4. also avoid it when the distribution is too unbalanced. for example, if one block has around 90 percent of the flow and the rest only 10 percent, then the chart will not be useful.

Image By Author

1. also do not use an alluvial diagram when a category has no change. when this happens, the diagram does not show anything meaningful.
2. for example, if in a country the age group 25 to 55 does not get government health benefits and the rest do, and this stays the same, then the alluvial diagram adds no real value. it becomes almost empty in terms of insight.

Image By Author

Time Gaps Can Break the Alluvial Flow

another problem is when the data has time gaps. let’s say one person completes a course in 3 years and another takes 4 years, even though both started from the same class.

in this case, the alluvial diagram will show gaps or uneven flows. this makes the analysis confusing because the timeline is not consistent.

do not confuse this with the previous example. earlier, the data had a complete missing connection. but here, the data exists, just with different durations.

this difference in time creates misalignment in the flow, and that can make the alluvial diagram harder to read and interpret.

Hey there, I’m Ajay a passionate engineer, writer on Medium, and a huge Python enthusiast. Thanks for sticking with me till the end!

If you enjoyed this content and want to support my work, the best ways are:

• Drop a Clap 👏 and share your thoughts 💬 below!
• Follow me here on Medium.
• Connect with me on LinkedIn.
• Follow me on X here.
• Join my email list so you never miss another article.
• And don’t forget to follow **The Visual Analyst** for more stories like this.

Thanks again for reading! 😊