The dumbbell chart is a slightly obscure variation of the line-chart/connected scatterplot. And is useful for comparing the change/difference between two categories such as date or between genders.
The dumbbell Chart is a two step process, first we assemble the base chart this is simply a line chart, here we have chosen the two end points of the line as the first and last year of the dataset, but this could be any breakdown, including discrete dimensions with two members.
Next we add the dimension that we want to compare, so for example here we are going to look at country:
Finally we add in the ‘weights’ to the end of the bars, this is done through a dual axis chart, so we duplicate one of the two axis, and then change the mark type on that axis;
Finally we make this dual axis, and then (if necessary) synchronize the two parts of the dual axis:
The obvious problem with dumbbell charts is that when we breakdown by a category with more than about five members we get a highly confusing mess as a graph;
There is another kind of dumbbell chart that is more suitable for comparing across dimensions with more members.
This is the dumbbell scatterplot:
Here the dumbbell looks at differences in gender across two categories, here increased length or ‘angle’ in the bar equates to bigger disagreement between the genders on these measures. This kind of dumbbell chart is built in much the same way as the standard dumbbell, but using a true measure on the x-axis and then deploying the small multiple ROW and COL calculations calculated over Country. (for more information on how to build a small multiple chart check out Andy Kriebel’s excellent blog post: https://goo.gl/vHZ32D ) you can check out the above chart over on my tableau public profile if you want to dissect how it was constructed: https://goo.gl/AY3kKF
As always let us know if you have any questions, and have fun with your dumbbell charts!