To start of: a dumbbell plot has nothing to do with being 'dumb'. On the contrary, it is a powerful representation of data, when you want to guide the viewer towards a visual discrepancy with a 'not so very often used' type of chart. In order to make it a bit more famous, let's check it out.
What the hell is a Dumbbell plot?
A dumbell plot has many names. Sometimes it is labelled as 'connected dotplot' or 'connected lollipop chart'. Knowing this makes it a bit more easier to understand that the name is derived from the 'fitness weight' you can find at the gym.
The two round edges at the end of the handle represent the value that you want to visualize. The handle itself ensures the connection between the two values, emphasizing the relation between the values.
A dumbbell (not-so-plot)
The main purpose of a dumbbell plot is to tell the story of comparison, where two values are compared with each other, grouped by a specified category.
Some examples:
- Paycheck difference between men and women in various countries
- Difference of sales between this and last year for various categories
- Correlation of inflation and unemployment rate througout the years
The two points are always connected with each other, which makes it easy to read and understand.
Below you can find an example, which I reworked starting from a bar chart billboard that I saw on vacation in an Italian zoo:
Reworked bar chart as a Dumbell chart:
Why a Dumbbell plot?
When reading what a dumbbell plot is, you would probably have guessed that a clustered bar chart would be sufficient as well for your reader, as they explain the same results. Almost. Due to the fact that the dots are connected, you, as a viewer are able to see the relation between the two values, whereas this is much more difficult in a clustered bar chart.
The only downside of this chart is that is it not used that often, and that experience is needed to first digest the visual elements. This could mean that your cognitive load of the chart is big at first, but the insight you get out of it is greater than the one you would get from a bar chart, trying to tell the same story
Dumbbell plot and BI tools, a match made in hell?
From my experience, when wanting to create a dumbbell plot, I really found it difficult to create it in traditional BI tools, like Power BI. It seems that this chart can only be created in tools like Tableau, where it still seems like a challenge.
So how do you create one? For this example I used a combination of Flourish and Canva (they are integrated) to visualize this data.
Conclusion
I really find dumbbell plots powerful visuals to explain a certain story when there is a big gap between two values based on a category. It of course depends on which story you want to tell. It seems that the chart is currently underused and that it is often swapped with the famous bar chart due to its familiarity, but nevertheless I feel that the dumbbell is here to stay.
Comments