Re: st: dynamic report generation

[Phil Schumm <pschumm@uchicago.edu>]

On Oct 27, 2006, at 12:55 PM, Jason Thompson wrote:
> is anyone using Stata for automatic document generation, ala Sweave  
> with R/S? i have discovered a presentation on reproducible research  
> here, http://repec.org/nasug2005/Schumm_NASUG-presentation.pdf, but  
> the command -stata2doc- does not seem to be available online. i  
> have an email in to the primary author but thought i'd also ask  
> around.

I am very sorry to say that since that presentation, little has been  
done on the project, and our efforts to secure funding for further  
development have so far been unsuccessful.  It is our intention to  
return to it at some point, but at the moment it is difficult to  
predict when that will happen.

At the time I gave that presentation we had made good progress on the  
code itself, but no documentation had been written (and that hasn't  
changed).  Moreover, no attempt had been made to package the system  
to make it easy to use; using it required an under-the-hood knowledge  
of both Python and docutils (in addition to Stata, of course).  I  
made no attempt to distribute it because at the time I knew I  
couldn't handle the inevitable support questions, and I didn't want  
to leave users (understandably) frustrated.

At the moment, the state of the art for automatic report generation  
is probably to use both -estout- and -graph- to write tables and  
figures to individual output files, and then to include these in the  
main document (e.g., using \include{} in LaTeX).  Alternatively, if  
you require more complex tables, you can write the table in LaTeX  
with empty cells, and then use Stata code to fill in the numbers.  I  
have used this approach for tasks such as generating quarterly data  
monitoring reports for a clinical trial.  Although it generates very  
nice results, it takes some practice in order to do it quickly.

Of course, neither of these approaches matches the level of  
integration of Sweave, nor, for example, are capable of placing  
individual results directly into the text.  I am pleased, however, to  
hear that there are others who are interested in this, and will  
notify you if (and when) we have something ready to distribute.


-- Phil

*
*   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/