« Sentiment analysis of Trump's tweets with R | Main | Because it's Friday: Gravity wave detecting lasers ... in space! »

August 19, 2016


Feed You can follow this conversation by subscribing to the comment feed for this post.

Thanks David for the article, I’m also in favour of selective approach to use dual axis in data analysis. In my case I had a dataset of 2 similar unit metrics but their values had a different range; and I tried to show this in my blog post: http://datanrg.blogspot.ca/2016/05/analyzing-david-and-goliath-datasets-on.html

Nice post. I would like to have two x axis and one y. Do not you know about solution?

Nice post, thanks for this! A commenter on my original post picked up an embarrassing error (although not material to the question at point) with the y axis label on my "fixed" chart. It should be "USD purchased with one NZD", not the other way around... I've fixed in the original, would you mind picking up a fresh copy so I don't have to wince if this is reproduced in future? http://ellisp.github.io/img/0051-dualgood.svg

Wait. You show five excellent reasons not to use dual axis plots and then proceed using exactly those?
That seems ironic. Am I missing something?

My take away is: multipanel plots are usually the best solution.
"will often be my preferred approach to this type of data".

Also, adding a third series is easy...

@Berry, check Peter's original post (last link in the above) for his excellent explanations of why dual-axis plots are usually bad, and the situations where they are useful. I usually prefer panel plots as well, but as Peter explains there are some downsides to those as well.

@Peter, you're welcome, and I've updated the chart above. Cheers!

The comments to this entry are closed.

Search Revolutions Blog

Got comments or suggestions for the blog editor?
Email David Smith.
Follow revodavid on Twitter Follow David on Twitter: @revodavid
Get this blog via email with Blogtrottr