Friday, May 31, 2013

Research Computing / Data / Writing Environments

You asked for favorite blogs, and I obliged. You also asked for favorite computing environments etc., so here goes. I look forward to comments telling me why I'm wrong, stupid, insane, or worse.

At some level, who cares about computing etc.? If you're deep into retirement, a SAS guru writing in WordPerfect (say), is it worth updating? Almost surely not.

But if your investment horizon is longer, and if you want to be on the cutting edge, and if you want my opinion, I certainly have one. What follows is in part prescriptive, although I realize that one size surely can't fit all. In any event it's certainly descriptive; it's basically what I do, in principle if not always in practice. (I admit that I'm still rather fond of certain ancient low-level environments like Fortran, and certain high-level environments like Eviews.)

My computing epiphany of recent years centers on R, a mid-level environment.  I find that R is most often the place to be.  Check R Studio IDE, a wonderful R work environment. Check R-bloggers (I should have mentioned it as a favorite blog -- thanks to Frank DiTraglia for reminding me). And for all you parallelization freaks with GPU's, check the CRAN Task View on High-Performance and Parallel Computing with R and the R Tutorial.

For time-series data, check Quandl. Totally amazing. Just click on the link and see for yourself. (And yes, there's a seamless R interface.) Imagine having basically any time-series you could ever want, instantly available and continuously updated, for use in your R code.

For writing, obviously it's LaTeX. My favorite flavor is MiKTeX. Enough said.

Now here's the first kicker. I already mentioned that Quandl and R are interfaced. But so too are R and LaTex, via Sweave. So now data, computing and writing are all linked. Imagine writing a book (in LaTeX) whose graphics and statistical analyses (in R) are automatically updated in real time as new data arrive (in Quandl). It's not a dream.

And here's the second kicker. Everything I've emphasized is public domain, open source, free. Who says that you get only what you pay for? This is highest quality everything, cutting edge, with no license hassles, no renewal hassles, no payment hassles.

Power to the people!