Attention R users, especially those that are experienced enough to be opinionated, I need your input. Consider the following simple plot: x <- rnorm(100) y <- rnorm(100) plot(x, y, bty='n') A colleague (and dreaded SAS user) commented that she thought that my plots could be "cleaned up" by connecting the X and Y axes. I know that I can do that with bty='l' but I don't want to, I find that the plots look less cluttered with disjoint axes. However, I was intrigued enough by her comments that I decided to solicit the opinions of others on this issue. Are there principled reasons why one should prefer joined axes or disjoint axes? Brant Inman
Okay, Brant. It's Friday and I'm opinionated. Since the plotted data extend to lower values on both axes that 'n' draws them, I like 'l' better. Clint Clint Bowman INTERNET: clint at ecy.wa.gov Air Dispersion Modeler INTERNET: clint at math.utah.edu Air Quality Program VOICE: (360) 407-6815 Department of Ecology FAX: (360) 407-7534 USPS: PO Box 47600, Olympia, WA 98504-7600 Parcels: 300 Desmond Drive, Lacey, WA 98503-1274 On Fri, 20 Apr 2007, Inman, Brant A. M.D. wrote:> > Attention R users, especially those that are experienced enough to be > opinionated, I need your input. > > Consider the following simple plot: > > x <- rnorm(100) > y <- rnorm(100) > plot(x, y, bty='n') > > A colleague (and dreaded SAS user) commented that she thought that my > plots could be "cleaned up" by connecting the X and Y axes. I know that > I can do that with bty='l' but I don't want to, I find that the plots > look less cluttered with disjoint axes. > > However, I was intrigued enough by her comments that I decided to > solicit the opinions of others on this issue. Are there principled > reasons why one should prefer joined axes or disjoint axes? > > Brant Inman > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. >
I would say ``de gustibus non disputandum'', except that I can't speak Latin. cheers, Rolf Turner rolf at math.unb.ca
Edward Tufte seems to have some opinions on this topic.
In The Visual Display of Quantitative Information (Chapter 6: Data-Ink
Maximization and Graphical Design -> Redesign of the Scatterplot), he
presents several alternatives
(1) "non-data-bearing frame" in conventional scatterplots (equivalent
to R's
bty="l"), which he argues is the common but less informative method.
(2) a little removal of ink from (1) can change axes to display the range of
the data ("range-frame").
(3) or, with slight modification of (2), fivenum().
(4) or even "dot-dash-plots" in which marginal frequency distribution
are
displayed as the axis using dots and dashes.
I don't know that bty="n" with xaxs,yaxs equal to "i" or
"r" meets any of
these criteria (and bty="l" is apparently less informative than his
other
suggestions)...
--- "Inman, Brant A. M.D." <Inman.Brant at mayo.edu> wrote:
>
> Attention R users, especially those that are experienced enough to be
> opinionated, I need your input.
>
> Consider the following simple plot:
>
> x <- rnorm(100)
> y <- rnorm(100)
> plot(x, y, bty='n')
>
> A colleague (and dreaded SAS user) commented that she thought that my
> plots could be "cleaned up" by connecting the X and Y axes. I
know that
> I can do that with bty='l' but I don't want to, I find that the
plots
> look less cluttered with disjoint axes.
>
> However, I was intrigued enough by her comments that I decided to
> solicit the opinions of others on this issue. Are there principled
> reasons why one should prefer joined axes or disjoint axes?
>
> Brant Inman
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>