Dual Axes: Suggested do’s and don’ts

A dual axes chart is used to visualize two measures on the same chart using two independent vertical axes (which can be synchronized).

Many articles describe how to use and misuse dual axes charts in data visualization, However, as a data visualization specialist and consultant, I have seen repeatedly misuse of dual axes, and in most cases, I find the recommendation of best practices to be a hard-sell.

This article will discuss the common misuses of dual axes charts and will suggest alternatives, as well as a quick video on how to do it in my go-to data visualization tool- Tableau.

Divide and conquer

In many cases, we wish to see the correlation between two biz related KPIs over time, such as Revenue, Profit Ratio, Order Qty. etc.
a common conception is to think that when displayed on the same graph, these related sets of values can be compared easily. Are they?

Let’s look at this example:

We wish to see the relationship between seemingly two measures that seemingly relate to the same unit- dollars.
but in fact, we have two different units (Dollars and %) and consequentially two very different scales.

Placing them on top of each other draws unnecessary clutter and makes our eyes wander between the different axes and lines, reducing clarity and increasing confusion.

Some alternatives I came across combines bar chart and line chart for the sake of creating a more distinct separation between the two:

IMO, this doesn’t improve the clarity and might even confuse the reader as bars tend to imply distinct measures for a specific point of time whereas line charts imply on continuous trend.

My suggestion for this use case is to create a distinct distinction for each one of the measures using a mutual x-axis for time, and a separate y-axis for each one of the measures:

Keep it simple: this way we can view and assess more clearly each one of the measures and also see how they relate to one another over time.

Moreover — we make sure we won’t cause biased correlations conceptions by our readers as a result of the different scales of the two axes. more on that in the next example

Compare Apples to Apples

Another misuse I often see is the combination of completely different units with a completely different scale on the same chart, such as Sales and order Quantity:

Many suggest this is a good practice when using a dual axes chart because again, this helps to create a more clear distinction between the two measures.
I disagree mainly because of the biased correlation perception it causes.

When we see two lines on top of each other our first instinct is often to look at the correlation between them. In this example, we see that in September the green line went over the grey line for the first time in 2022.
Then we see that the green line is the quantity and the grey line sales. So in September Quantity is bigger than sales right? But wait, these two measures aren’t really comparable aren’t they?!

Another potential bias is the magnitude of the change or the angle of the slope in our case. Different scales generate different angles and they are not comparable!

If correlation over time between 2 measures is the data story we wish to convey, we have to make sure we compare apples to apples.
To achieve that we can bring the two measures to a common ground and track the percentage of changes from it:

Look how different the correlation and slopes are from previous chart example!

Here is a quick video of how to visualize it in Tableau:

To wrap this up:

Avoid mixing different scales and units on the same chart. Avoid clutter and think of how you can provide the maximum unbiased clarity to your readers.

Inspired by this highly recommended article by Stephen Few.

Link to downloadable Tableau workbook here

Thanks for reading, share your thoughts!