[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: RE: scatterplot matrix with predicted plots
Garry wanted an option to -graph matrix-.
I was explaining my guess that this isn't very easy.
Also, I got the impression that he is not a Stata
programmer, and so was not interested
in answers of the form "try writing your own program".
Yes, you can try fudging this for yourself. The
code you quote shows one approach. I pushed it
a bit further after that thread on Statalist.
Vince Wiggins gave valuable assists. However,
although I didn't say so on Statalist, I came
to the conclusion that I couldn't develop it
into a program I thought worth making fully public (i.e.
posted on SSC with associated help). It
needed too much understanding of undocumented
features, and especially more time than I
wanted to give the problem.
> -----Original Message-----
> From: email@example.com
> [mailto:firstname.lastname@example.org]On Behalf Of
> Roger Harbord
> Sent: 26 May 2004 19:59
> To: email@example.com
> Subject: Re: st: RE: scatterplot matrix with predicted plots
> Didn't you post code to statalist to do something very
> similar to this just
> under a year ago?? I know you post an amazing amount of code
> to statalist but
> i'm surprised if you've forgotten...
> I have this in my personal ado directory:
> . which mygraphmatrix
> *! NJC 1.0.0 12 June 2003
> *! Posted on Statalist.
> *! Matrix plot with lowess smooths and quantile plots on diagonal.
> *! It's slow.
> Granted it's kind of a brute-force approach - it draws each
> graph and saves it
> then -graph combine-s the results, so you get axes around
> each sub-graph and
> it isn't as efficient in use of time or space as a low-level
> extension to
> -graph matrix- possibly could be. It's still a very nice
> tool for data
> exploration though (with a built-in an excuse for a quick tea break).
> Garry, if you search the Statalist archives (
> http://www.stata.com/statalist/archive/ ) for -mygraphmatrix-
> you'll find the
> code. Should be easy enough to adapt to use other fits
> instead of lowess if
> desired (might run a little quicker too).
> --On 26 May 2004 13:29 +0100 Nick Cox <firstname.lastname@example.org> wrote:
> > This is a very understandable thing to
> > want. My guess, however, is that it
> > is much more difficult to implement
> > than you would guess. That is, it
> > would be possible to add code at the
> > low level that would let you slap
> > something Y X
> > on top of
> > scatter Y X
> > The difficulty is to do that while
> > also giving you, the user, the kinds
> > of handles that you really want
> > to control the details of the graphs.
> > Thus you might say, "All I want is an lfit Y X
> > on top on each scatter." But if that
> > is done blindly, then not everything
> > is what you really want: for example,
> > the y axis titles would probably be
> > too messy for your taste. So you
> > need a handle that controls that.
> > All this is guesswork, and a signal
> > to StataCorp to say that I am way
> > too pessimistic here. We're in weird
> > territory with the new graphics:
> > some things which might sound very difficult
> > are in fact very easy, and vice versa too.
> > Nick
> > email@example.com
> >> -----Original Message-----
> >> From: firstname.lastname@example.org
> >> [mailto:email@example.com]On Behalf Of
> >> Garry Anderson
> >> Sent: 26 May 2004 08:13
> >> To: firstname.lastname@example.org
> >> Subject: st: scatterplot matrix with predicted plots
> >> Hi,
> >> Is it possible please to overlay a predicted plot (linear or
> >> quadratic or
> >> fractional polynomial etc) for each graph in a scatterplot matrix?
> >> There does not appear to be an option in
> >> graph matrix x1 x2 x3 x4 y
> >> or in the menu for scatterplot matrix
> >> Best wishes, Garry
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
* For searches and help try: