Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: RE: Stata and R


From   David Airey <david.airey@Vanderbilt.Edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: RE: Stata and R
Date   Fri, 16 May 2008 13:27:33 -0500

.

I think you are right. In trying out various R code examples that make use of Design's nomogram() function (even in the help file for this command), there seems to be difficulty in actually making the graph look good. There are problems in relative size of the scales in terms of readability, problems in size of variable labels, problems in the size of the optimal graphics device aspect ratio, etc. That said, when the graph looks good, it is an outstanding graphic device for presenting a fitted model to a non-statistician. But, as you say, it is not an easy program to knock off, and it is questionable how useful it really is. Cool, but enough so to replace or complement estout?


On May 16, 2008, at 1:01 PM, Nick Cox wrote:


A first impression is that it is very possible as a Stata program, but
quite tricky to do well and with some generality. That reduces the
chance that some programmer would knock it off as a Sudoku-like puzzle.
You would have to want to be able to use it, very much. I would be
pleased to be shown wrong on that first impression.

Nick
n.j.cox@durham.ac.uk

David Airey

I followed this link too, and while it is a nomogram (Fagan's
nomogram) it is not what I'm after. Nomogram() in R is more general
and follows any GLM. See the example picture in the first link I gave.

-Dave

On May 15, 2008, at 4:11 PM, Nick Cox wrote:


-findit nomogram- points to Ben Dwamena's -midas-.

Nick
n.j.cox@durham.ac.uk

David Airey

Just about every other week I drop into our R clinic for a coffee
break to listen to people's questions and maybe ask one myself. The
clinic is about general use questions, not especially about
Bioconductor packages, which is one reason to use R. The clinic is run
by a very capable R expert. Originally I started going to the clinic
thinking to myself that well, maybe this would be the last version of
Stata I would buy, because R has come miles in development and
documentation. More often than not, however, I conclude the opposite,
and continue to regularly use Stata, thinking there is value is good
commercial products.

One interesting function in R that I can't find in Stata is Frank
Harrell's nomogram function in the Design package, which essentially
is a graphical alternative to outreg or estout.

http://lib.stat.cmu.edu/S/Harrell/help/Design/html/nomogram.html


http://www.stat.columbia.edu/~cook/movabletype/archives/2006/05/nomogram
s.html
*
*   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:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index