Statisfaction

Running R on an iPhone/iPad with RStudio

Posted in R by Julyan Arbel on 3 May 2011

This thread has been widely discussed on a lot of forums. To make a long story short, running natively R on an iDevice (meaning iPhone/iPad) is disabled by its OS, unless it is jailbroken. The steps for the installation through Cydia are described in this R wiki, or this post. But there are some limitations, including bugs in package management. Running R directly seems not to be the more sensible way to proceed: smartphones lack so much power for that, you would do it just for fun. Typing R code on a small keyboard is quite tedious. But is manageable on a larger one (iPad, Galaxy Tab).

There could be another short-term solution to use R on an iDevice (or any other smartphone). As explained on this introduction to RStudio, the novel IDE can be launched directly on the Web, while running on a server. This would be an easier and safer solution than jailbreaking the mobile, which would have a number of benefits, including: the ability to access and share your files from anywhere (with an Internet connection), to benefit from more powerful computing resources (of your server), and to focus on a single configuration of R (for packages).

But there is a but, because the RStudio development team is still working on it. It is close to working (it is possible to click on tabs and menus) but not yet possible to get the keyboard to initialize and to type in the console. Yet executing the code can be done by clicking on Run line(s)+Enter. It is not a lot for now, but is a good start.

Tagged with: , , ,

RStudio is good for you

Posted in R by Julyan Arbel on 29 April 2011

I was recently introduced to RStudio, a new integrated development environment for R, it is just amazing!
It is free, and open, compatible with PC/Mac/Linux OSs. You can also choose to run it in the cloud, and access it from your favorite web browser.

As you can see, the window divides into four in a customizable way. For the source code, multiple files are organized by tabs. The editor is great, providing code completion (for commands and existing variables), indentation, and syntax highlighting. Execution from source can be done with the Run icons (as well as saving, printing), or by a shortcut (Command or Ctrl+Enter). For more shortcuts you can check this page.

Another quadrant is for Workspace and History. The first lists as in Matlab all your current variables. You can have a quick look to them by clicking (equivalent to the View command), and edit your functions (equivalent to the fix command). It makes easy to import datasets from text files or from the Web. The bottom left quadrant allows to browse your files. A very interesting feat is the Plots tab, which stores all the past plots. You can export at the end the one you want, in pdf or png. It includes as well a package installation tool, and the R documentation.

For enabling access from any Web browser, there exists the RStudio Server version, to be installed on a Linux server. It’s great for team projects, as you can enable access to anyone by giving the server url. It allows to install exotic packages on a single server, which is good when you have restrictions on other computers. And you can check from a remote position the results of your computations you are drawing for the week-end. It looks like that:

For people at ENSAE, admin registration is not required for installing RStudio, so don’t hesitate to give it a try!

Tagged with: , ,
%d bloggers like this: