Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: R for Stata Users
Stas Kolenikov <email@example.com>
Re: st: R for Stata Users
Fri, 26 Feb 2010 09:35:55 -0600
On Fri, Feb 26, 2010 at 8:42 AM, Airey, David C
> Rather I might want to know how to use R to extend Stata with packages
> unavailable in Stata and outside my expertise to do so, period.
+1. Moreover, I think in the long run there will be more instances of lousy
analysis done using R that using Stata. A lot of R users come to use the
package based on the professional programmers' excitement about R as a
programming environment, but producing an efficient (and accurate) data
analysis requires relatively deep knowledge of both what you are doing in
terms of statistics, and how exactly you do this in R.
I admire R graphics for its speed and the ability to produce an endless
range of contour plots... except that I need to spend half an hour
formatting my data to get the grid the R packages need. I like R's ability
to produce graphs directly from matrices -- in Mata, you'd have to
-st_store- things first somehow to call your -twoway- or whatever from
Stata. (Or is there a better way?) I like R's option management -- you can
transfer options by position or by name, but I don't know how much you can
do in terms of suboptions with complicated parsing. And that's where my R
knowledge stops. I am still lost at the differences between lists, data
frames, matrices and whatever other objects R might have (and my programming
background is somewhat above lousy, I think, so I should be able to grasp
these concepts if they were clearly explained SOMEWHERE). I know that a
rough analogue of -by- is *apply, but I don't quite know what the analogies
are. I've no clue about data manipulation whatsoever (except that rbind() is
-append- and cbind() is -merge- provided you sorted everything yourself). I
know that -if- modifier is surprisingly difficult to implement in R, so
you'd have to use all of your data set most of the time. If I had a book
that would explain those points, that would be great. But as I am yet to see
such book, I use R by and large as a graphics calculator rather than a
statistical package or a programming environment. Stata gave me enough
programming tools to write pretty much anything I need for my research
purposes. But since I am neither a typical Stata user nor a typical R user,
all of the above is my humble opinion.
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
* For searches and help try: