Re: st: Popularity of R, SAS, SPSS, Stata, Statistica, S-PLUS updated

[Christopher Baum <kit.baum@bc.edu>]

Hi Bob,

Those stats are only available monthly. The stats accessed by -ssc hot- within Stata are available quite a bit farther back, but they are not counts of raw downloads. In any case what is being counted on report.ssc.html is the download of individual files. Many packages (the equivalent of R packages) contain a number of files (all contain at least 2, but some contain many more). Thus the stats accessible within Stata are counts of package downloads,not file downloads. See my posting near the beginning of each month (except this month, when it was later) on SSC Activity for the top 25 package counts.

I can count how many packages are on SSC (1,407 at this moment) but I do not know how many are on STB/SJ, UCLA, etc. In any case such counts are overlapping, as often a package shows up first on SSC, then appears in the SJ, then is updated on SSC (and possibly on SJ). So there is some amount of duplication between SSC and SJ/STB. In some cases packages on SSC become redundant for users of the latest version of Stata, as it may contain an official command with that functionality. But I would say the vast majority of public Stata packages are either in SSC and/or STB/SJ archives.

Cheers
Kit

On Mar 25, 2011, at 9:52 AM, Muenchen, Robert A (Bob) wrote:

> That's very helpful information. I've switched the URL to point to:
> http://fmwww.bc.edu/fmrc/reports/report.ssc.html 
> 
> That link has a rich set of information regarding downloads: hourly, daily, weekly.  Is yearly download data available? If it is, it would make a great addition to the paper. 
> 
> Are there counts of the total packages there now? By year?  
> 
> Does anyone know what percent of the total of the main repositories are at SSC? To find equivalent info for R I chose only the biggest, CRAN, and ran a program to count the unique package names (2,849 on 3/25). I then selected all 9 major repositories and ran it again (4,338). So while yearly total number of R packages is known only for CRAN, we can estimate that the growth curve shown in Figure 9 (http://r4stats.com/popularity) is 66% of the total. Individuals still have their own sets, but probably a relatively small number. I would *love* to have similar data for Stata!
> 
> Thanks,
> Bob
> 
>> -----Original Message-----
>> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-
>> statalist@hsphsun2.harvard.edu] On Behalf Of Christopher Baum
>> Sent: Wednesday, March 23, 2011 1:27 PM
>> To: statalist@hsphsun2.harvard.edu
>> Subject: re: Re: st: Popularity of R, SAS, SPSS, Stata, Statistica, S-PLUS updated
>> 
>> <>
>> Nick Cox pointed out, in re "Popularity",
>> 
>> In this document there is a repeated misunderstanding. Downloads from SSC
>> (repec), as members of this list will generally know, are only one way of
>> downloading user-written packages for Stata. Indeed another main way, through
>> the Stata Journal and Stata Technical Bulletin websites, predates SSC. Data for
>> such downloads are proprietary to StataCorp. It's my guess that they are easily
>> the same order of magnitude as SSC downloads.
>> 
>> 
>> Actually, it is worse than that. The Muenchen document states
>> 
>> Similar figures for downloads of Stata add-ons (not Stata itself) are available at
>> http://logec.repec.org/scripts/itemstat.pf?type=redif-software.  However, both
>> R and Stata have other significant repositories that do not provide such counts.
>> 
>> This is quite misleading. The LogEc count referred to by that URL is the count of
>> .ado, .hlp, .sthlp, .mlib files downloaded from the SSC Archive via web browser
>> links. On each web browser page, users are strongly encouraged NOT to
>> download software this way (and that it will probably fail if they are Windows
>> users). Nevertheless, a nontrivial number do so; this statistic might be termed the
>> 'bozo count'. But those counts are a serious underestimate of download activity
>> from the SSC Archive, as the recommended technique to perform these
>> downloads is from within Stata using the -ssc- command, and a much larger
>> number of Stata users download using this recommendation. Such downloads are
>> not tracked by RePEc services, but they are tracked by the web server which
>> satisfies Stata's requests. So the URL above should be removed from the
>> "Popularity" document. A URL which gives a snapshot of download activity over a
>> recent month is
>> 
>> http://fmwww.bc.edu/fmrc/reports/report.ssc.html
>> 
>> Much more readable stats are available from within Stata, using -ssc hot-, but
>> that of course requires that you use Stata.
>> 
>> Nick's point -- that downloads of user-written software involve the SJ/STB
>> Archive, the UCLA archive and various users' sites -- is of course well taken.
>> 
>> Kit
>> 
>> Kit Baum   |   Boston College Economics and DIW Berlin   |
>> http://ideas.repec.org/e/pba1.html
>> An Introduction to Stata Programming   |   http://www.stata-
>> press.com/books/isp.html
>> An Introduction to Modern Econometrics Using Stata   |   http://www.stata-
>> press.com/books/imeus.html
>> 
>> 
>> *
>> *   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/
> 
> 


Research Papers in Economics (RePEc)
http://repec.org





*
*   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/