[fusion_builder_container hundred_percent=”yes” overflow=”visible”][fusion_builder_row][fusion_builder_column type=”1_1″ background_position=”left top” background_color=”” border_size=”” border_color=”” border_style=”solid” spacing=”yes” background_image=”” background_repeat=”no-repeat” padding=”” margin_top=”0px” margin_bottom=”0px” class=”” id=”” animation_type=”” animation_speed=”0.3″ animation_direction=”left” hide_on_mobile=”no” center_content=”no” min_height=”none”]
Here is an interesting technique: using two y-axis to display the same data at different resolutions. Yellow (BMI) and blue (Weight) lines should overlap (there is a one-to-one correspondence between BMI and Weight for a given height), but they don’t because they are using a higher resolution for Y2. That’s clever. We can choose to see more detail (BMI) or less detail (kg).
Unless…
Unless they don’t know what they are doing. There are many reasons why you shouldn’t use a secondary y-axis in your charts and this chart proves it. You can use the Y2 axis:
- to replicate Y1 (it helps reading the chart);
- to show one-to-one correspondences (Celsius/Fahrenheit, km/miles)
- Everything else is potentially manipulative.
You must make sure those correspondences are preserved. That’s not the case here: 26.5 is not the right BMI for 74 kg (for this height). If I remove the green line, for the same BMI I get 82.3 kg. So, different correspondences for different resolutions. Hence the title of this post.
I wanted to focus on this, but there are other issues (terrible smoothing algorithm; is the green line (Goal) assigned to Y1 or Y2? You can never know).
How do you use the secondary y-axis?
[/fusion_builder_column][/fusion_builder_row][/fusion_builder_container]
Agreed, great caution should be taken when adding a second Y axis. It easily distorts the data and the message as in the example.
Replication can be good, if you have timeseries with a long span. Besides that, Y2 is perfect for controling dummy series in the chart (e.g. positioning of labels or height of background “stripes”). But then, of course, the axis is hidden…
Secondary axes can be evil. I wrote about this in Secondary Axes in Charts, which in fact was a follow up to Stephen Few’s Dual-Scaled Axes in Graphs-Are They Ever the Best Solution?.
In my article I presented a few techniques to avoid the cognitive perils of secondary axes, including the use of normalized data and of panel charts.
That said, I’m a huge user of secondary axes in Excel charts. As Ulrik points out, they are indispensable when using hidden series for precise positioning of labels and shapes.
Jon, thanks for pointing to those articles.
Jon, Ulrik, the secondary axis is invaluable to go beyond the Excel chart gallery, but that’s not what a regular user uses it for…
When both axes start from zero (which is something i see as crucial except for interval scales), i use a secondary axis to show connected features, when both the total and the per capita figure seem important. for example total domestic water consumption and total population (from the jerusalem statistical yearbook)