Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Nick Cox <n.j.cox@durham.ac.uk> |
To | "'statalist@hsphsun2.harvard.edu'" <statalist@hsphsun2.harvard.edu> |
Subject | RE: st: Re: Problem with -graph bar- and yscale(log) option |
Date | Wed, 5 Oct 2011 16:40:43 +0100 |
I agree broadly with Maarten. It is difficult for me to see that bar charts mix with log scales at all. The key idea is naturally that bars encode magnitudes relative to some origin. If there is no origin, relative bar lengths just depend on some arbitrary choices about axis limits. On the other hand, this is precisely where -graph dot- can be useful. Nick n.j.cox@durham.ac.uk Maarten Buis Such bar graphs by default include 0 on the y-axis, which makes sense for such graphs but does not have an equivalent on a log-scale leading to the problem you noticed. You can avoid 0 by adding the -exclude0- option, but now it is up to you to decide whether that graph still makes sense. On Wed, Oct 5, 2011 at 3:23 PM, Aleksander Rutkowski wrote: > I am using Stata/SE 11.2 on Windows. I have noticed that using -graph > bar- with yscale(log) option produces very bad results: bars of equal > height with unreadable y-axis labels (overlaid at the same point). As > an illustration: > > sysuse auto > graph bar (mean) trunk, over(foreign) yscale(log) > > Is it a bug or am I doing something wrong? * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/