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]
st: re: speed of R graphics vs Stata graphics
From
"Airey, David C" <[email protected]>
To
"[email protected]" <[email protected]>
Subject
st: re: speed of R graphics vs Stata graphics
Date
Wed, 1 Sep 2010 18:53:32 -0500
.
> I was cutting my teeth on some graphics related analysis of gene microarrays using R Bioconductor packages. In some of the plots there may be 10K to 40K data points. I noticed a significant speed advantage to the R graphics in such plots with many symbols, and with trellis (panel) graphics. To confirm my impression I plotted 100K points generated using corr2data with a correlation of 0.5 in Stata and R using scatter y x and with(mydata(plot(x,y)). Stata draws each symbol as you watch and took 9 seconds. R draws the plot off screen and then it appears after less than 3 seconds. I'm using StataMP 11.1 and R 2.11.1.
>
> (1) Anybody notice this speed advantage to R with graphics?
>
> (2) Anybody notice Stata and R about both have 11.1 in their version numbers?
>
>
Now this is fascinating:
. scatter y x, ms(oh) aspect(1)
r; t=8.57 18:45:15
. scatter y x, ms(oh) aspect(1) mlw(vthin)
r; t=8.45 18:46:14
. scatter y x, ms(oh) aspect(1) mlw(vvthin)
r; t=7.73 18:48:59
. scatter y x, ms(point) aspect(1)
r; t=6.21 18:49:44
It takes less time in Stata to graph smaller symbols or symbols with less paint!
*
* 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/