Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 2 down vote favorite

I have a grid of plots (see example below) with different x axes but all the same y-axis. What is the best way to label the y-axis without being redundant? The information is in the legend, but this is not a solution.

Here are two options I am considering:

  • center one label on the right hand side in the same font as the x-axis labels. This could make it ambiguous if the y-axis label is associated with only one subplot.
  • leave the y-axis empty and provide information in a figure title (response of 'y-axis label (units)' to 15 variables).

This plot will be published in a scientific journal. Low resolution is intentional for this post since it is unpubliished. Any other suggestions appreciated.

enter image description here

share|improve this question
I do not see any problem with this, as long as the label is not pneumonoultramicroscopicsilicovolcanoconiosis. You can find many examples like this in AoAS. I would be more worried about the grid not being too bold. Another option is to specify the meaning of the $y-$axis in the caption. – user10525 May 23 '12 at 18:45
1  
@whuber thanks for the examples, they are more clear than I had imagined. – David May 23 '12 at 21:19
1  
In case you want more, I found them with a Google image search for "R trellis graphics". You have to be choosy, though: evidently, most statistical graphics one finds on the Web are awful :-). – whuber♦ May 23 '12 at 21:23
1  
@whuber I also found quite a few copies flipping through a stack of actual bound journals that ended up in the recycling bin. The y-axis label seems to do fine, and it is nice not to have to read the legend. – David May 23 '12 at 21:42
1  
@Jake It is not clear that it would make the plots easier to digest. My assumption is that it would takes up space and be distracting. If I saw sixteen y-axis labels, I would assume that they did not all contain the same information. – David May 23 '12 at 22:55
show 4 more comments

1 Answer

up vote 1 down vote accepted

I don't know what the perfect solution is here. I might ask others in the same field, or who have experience with that journal what has worked for them or what they suggest. My initial thought would be to label y-axes on the left hand side of the plots that are in the leftmost column. Then I might decrease the vertical margins between the plots relative to the horizontal margins between the plots. The gestalt will make the rows salient and visually highlight that they all share the same y-axis labels on the left. (You'd have to see how this looks, but I think it might work.) Furthermore, I would group the plots with the same x-axes by column if possible (even though I gather you're still going to have x-axes for every plot), and definitely point these facts out in the figure caption.

I should say, at this point, that I can't read your figure, I wonder if you could make and submit a high-resolution version and let the typesetting people reduce the resolution or figure out what they want to do with it. For example, sometimes they can keep a higher-resolution version at the journal's website or something.

share|improve this answer
thanks for your point... I can submit a higher resolution version after it is published. – David Jun 5 '12 at 16:36
2  
Just a note regarding the submission of high-resolution figures for publication: I believe that charts should always be submitted as vector graphics, not as raster graphics, so questions of resolution are not relevant. Publishing fuzzy plots may cause readers to cast doubt on otherwise outstanding research. I sometimes wonder if some researchers simply crop screenshots of their Excel workspace. – jthetzel Jun 5 '12 at 16:40
@jthetzel, good point. – gung Jun 5 '12 at 16:42

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.