References supporting the default aspect of ggplot2 graphics

[JiHO]

Hello everyone,

I just received back a manuscript in which I used ggplot2 exclusively
for the graphics (and it was kind of accepted, yay!). I tried to
follow the guidelines of the journal (fonts, dimensions, no color,
etc.) while retaining some of what make ggplot2 look and work great
(grey background, grid lines, legends outside the plotting panel,
etc.). In the edited copy, the editorial office basically recommends
that I remove all that. The book underlines that the default choices
of ggplot2 are there for a reason and, while I will conform to the
request of the editorial office, I would also like to point why the
plots where formated that way in the first place. My secret hope is
that the person in the editorial office who proofread the manuscript
was just part of the technical staff and applied the recommendations
without giving them any thoughts. Maybe someone higher up (the
editor?) could take the decision to accept the look of ggplot graphics
if it is correctly justified.

So here is my question: could someone (Hadley?) please point me
towards academic references that justify the defaults for ggplots (in
particular the grey background, the grid lines and the formatting of
legends)? I saw the references to the books by Tufte (1990, 1997,
2001, 2006) in the ggplot book, but I am looking for something more
precise (or at least some page ranges in those books since I don't
have access to any of these in my current position). And having
several authors concur on the subject would help because, even though
Tufte is considered the reference in this domain, the editor of the
journal I am writing for is not likely to know his work and could see
this as just one man's opinion.

Thanks very much in advance,

JiHO
---
http://maururu.net

[Mike Lawrence]

A start (on gridlines) comes from a previous post by Hadley (thread:
http://groups.google.com/group/ggplot2/browse_thread/thread/c0e47f343e119f7):

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: hadley wickham <h.wick...@gmail.com>
Date: Thu, 2 Jul 2009 09:06:12 -0500
Local: Thurs, Jul 2 2009 11:06 am
Subject: Re: Getting rid of horizontal grid lines
On Fri, Jun 26, 2009 at 12:41 PM, Peter

Flom<peterflomconsult...@mindspring.com> wrote:
> The whole question gets interesting.
> How much of the lines are what Tufte would call 'chartjunk'? I am sure Tufte would
> prefer that graphs contain a lot fewer gridlines than most ggplot graphs contain. I am
> also sure that I think Tufte goes too far .... his aesthetic purity of "bits of info per square inch"
> sacrifices some clarity. It's not how much info you can cram onto a page, it's how much the reader
> can get out, and how easily.

I agree. See the following paper for an example of why you shouldn't
accept Tufte's advice uncritically:

W. A. Stock and J. T. Behrens. Box, line, and midgap plots: Effects of
display characteristics on the accuracy and bias of estimates of
whisker length. Journal of Educational Statistics, 16(1): 1–20, 1991.

And section 3 of the following paper gives some concrete reasons why
you actually want grid lines:

W. Cleveland. A model for studying display methods of statistical
graphics. Journal of Computational and Graphical Statistics, 2:
323–364, 1993. URL http://stat.bell-labs.com/doc/93.4.ps.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

[hadley wickham]

> So here is my question: could someone (Hadley?) please point me
> towards academic references that justify the defaults for ggplots (in
> particular the grey background, the grid lines and the formatting of
> legends)? I saw the references to the books by Tufte (1990, 1997,
> 2001, 2006) in the ggplot book, but I am looking for something more
> precise (or at least some page ranges in those books since I don't
> have access to any of these in my current position). And having
> several authors concur on the subject would help because, even though
> Tufte is considered the reference in this domain, the editor of the
> journal I am writing for is not likely to know his work and could see
> this as just one man's opinion.

There's not a lot, unfortunately. Dan Carr had a couple of nice
articles in the SCGN newsletter:

D. Carr. Using gray in plots. ASA Statistical Computing and Graphics
Newsletter, 2 (5):11–14, 1994. URL
http://www.galaxy.gmu.edu/~dcarr/lib/v5n2.pdf.

D. Carr and R. Sun. Using layering and perceptual grouping in
statistical graphics. ASA Statistical Computing and Graphics
Newsletter, 10 (1):25–31, 1999.

He also has some good general advice in the following article, but I'm
not sure if he specifically mentions gray backgrounds (although he
does use them a lot)

D. Carr. Graphical displays. In A. H. El-Shaarawi and W. W. Piegorsch,
editors, Encyclopedia of Environmetrics, volume 2, pages 933–960. John
Wiley & Sons, 2002. URL
http://www.galaxy.gmu.edu/~dcarr/lib/EnvironmentalGraphics.pdf.

Hadley

--
http://had.co.nz/

[JiHO]

Thanks for the replies Mike and Hadley. I swear I did a search for
post regarding gridlines on the ggplot2 archive on google and it
turned up nothing. I guess I searched for "grid lines", in two words.
Anyhow, I'll compile those refs, try to find some of the articles and
write up my argument to the editor. I'll let you know how it turns out.

Thanks again. Sincerely,

JiHO
---
http://maururu.net