Really enjoying RStudio‘s Shiny Developer Conference | Stanford University | January 2016.
Winston Chang just demonstrated profvis, really slick. You can profile code just by wrapping it in a profvis({}) block and the results are exported as interactive HTML widgets.
For example, running the R code below:
if(!('profvis' %in% rownames(installed.packages()))) {
devtools::install_github('rstudio/profvis')
}
library('profvis')
nrow = 10000
ncol = 1000
data <- as.data.frame(matrix(rnorm(nrow*ncol),
nrow=nrow,ncol=ncol))
profvis({
d <- data
means <- apply(d,2,mean)
for(i in seq_along(means)) {
d[[i]] <- d[[i]] - means[[i]]
}
})
Produces an interactive version of the following profile information:
Definitely check it out!
Many other great presentations, this one is just particularly easy to share.
We will be bringing a lot of techniques back and to our clients.
Categories: Exciting Techniques Tutorials
jmount
Data Scientist and trainer at Win Vector LLC. One of the authors of Practical Data Science with R.
It might be a very good tool for profiling R code but every time I try to run my code RStudio crashed and start a new session.
I have found the package to be a bit install dependent. It worked out of the box on my MacBook Air (which I posted this article from, during the conference), but to get it to work on my office Mac Mini I needed to do the following at the command line (without instructions):
John, I enjoyed that talk too. It was also nice to meet you in person!