[R] R Reporting - PDF/HTML mature presentation quality package?

Philippe Grosjean phgrosjean at sciviews.org
Sun Jun 25 10:36:32 CEST 2006


Frank E Harrell Jr wrote:
> zubin wrote:
> 
>>Hello, searching for the key packages so i can output, Text, Tables ,and 
>>Graphics into a HTML or PDF report - have R create these reports in an 
>>easy and efficient way without LaTeX - I have searched the R pages but 
>>don't see any mature packages - anyone have any advice on a easy to use 
>>R package one can use for generating publication quality reports?  
>>Outputting HTML or PDF.
> 
> 
> Doing this without LaTeX is like doing statistical analysis without 
> linear models and the Wilcoxon test.
> 

It depends the complexity of formatting and the intended use of the 
report. If required formatting is just title, bold, italic, etc. + 
tables and graphs inserted in the same page to be viewed in a web 
browser, HTML (or xHTML/CSS, to be more precise) can do a pretty good 
job (and it can do even much more). Of course, LaTeX + Sweave => PDF is 
another solution, richer in functionnalities, but more complex to 
install, especially on Windows.

Zubin, you ask for "mature package". You should look at R2HTML + svViews 
(from the SciViews bundle) that work hand in hand to generate chunks of 
xHTML/CSS elements from R objects (the so called, views), or complete 
HTML page collecting together several of these views (reports). 
Depending on the "degree of maturity" that is acceptable for you, you 
should perhaps plan a beta test and expect some code adaptation, as it 
is usual with any R package that is not intensively used.

If you prefer dynamic reports, there are many approaches, but I tend to 
consider Rpad (also on CRAN) as one of the better solution out there. 
Note that it also uses R2HTML for R objects formatting, although Tom 
Short customized some of the R2HTML functions for a better use in Rpad. 
You should probably also customize and/or contribute to the development 
of R2HTML if you want a solution that best fits your own needs in any case.

Best,

Philippe Grosjean



More information about the R-help mailing list