[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: RE: RE: RE: vwidth in old boxplot graphics
I don't think these comments are a totally fair summary
of the situation, quite apart from the rather
strange comment "(especially important with the
non-statistical biomedical crowd)". Exploration and
visualization are important for everyone who uses
data, biomedical or not, statistical or not.
Naturally you are entitled to your perspective, but
My reading of the runes is that -vwidth- will not
come ever back within official Stata. I guess there's
a mix of reasons for that, some computational,
some statistical, some aesthetic. So that is bad news for those
like you who liked it. So you are right that you will need to
fire up -graph7- for that.
I've looked at mosaic plots, and I guess they're fairly
easily programmed in Stata. I have not been tempted to
do that myself, but someone else might. (The related but
much simpler notions implemented in my -tabplot- and -tableplot-
I find helpful for two-way tables.) But of course
mosaic plots were not available in Stata 7 either.
However, although there have been a few casualties like
-vwidth-, the new graphics is so much flexible and versatile
than the old that I couldn't agree less with your (apparent) sweeping
dismissal of it.
> Thanks for the prompt and detailed reply. Evidently most things have
> gotten better as versions progress whereas some have gotten more
> difficult. Stata's advantages, among others, is its ability
> to quickly
> explore and visually highlight data (especially important with the
> non-statistical biomedical crowd). Perhaps subsequent versions will
> reinstate features such as [graph y box, by(x) vwidth] or
> similar things
> for categorical data, such as mosaic plots (see
> http://www.math.yorku.ca/SCS/friendly.html), making crosstabulations
> more readable. Till then I guess its down to using Stata 7.0
> and to do
> without fancy graphics.
* For searches and help try: