Skip to content

Latest commit

 

History

History
94 lines (53 loc) · 4.93 KB

vm_inst.md

File metadata and controls

94 lines (53 loc) · 4.93 KB
Raw

One option for running R and RStudio is to use a "Virtual Machine" VM, hosted by UO IT and CASIT. A virtual machine runs a server, and is accessed (in this case) through a web browser. The VMs that have been "provisioned" for us run R and RStudio only a little more slowly than a "local" machine, and could provide a workaround if your access to a newish computer is limited. The instructions here cover the initial setup that needs to be done in order to insure that your workspace and data are saved between sessions.

Setting up the virtual machine

Logon/start the VM

To start:

  • Open Firefox, and click on this URL: https://view.uoregon.edu/
  • Login using your UO userid (the part of your email address in front of the @ sign) and password;
  • Click (or double-click) on the UO Virtual Computer Lab icon;
  • respond Yes to the cut-and-paste option.

A standard Windows 10 boot-up screen should appear. It takes a little while for the virtual machine to start up. Give everything a little more time after clicking on something than you would expect on a laptop or lab machine.

Start RStudio

To start RStudio,

  • click on the Windows Start button, and
  • type "RStudio" (without the quotes); then
  • click on the RStudio Desktop app bar

RStudio will start in a few seconds (wait). Then (this is important), in the Console window of RStudio, type the following, replacing "userid" with your login/user id:

 setwd("R:/geog495_1/Student_Data/userid/") 
 .libPaths("R:/geog495_1/Class_Data/win-library/3.6")

Note the forward slashes in the paths (the backward slash in R is an escape character), and make sure the cases (upper or lower) of the characters are strictly followed. The first command sets the "working directroy" (where R saves data and plots) to the Class_Data folder, while the second indicates where downloaded packages should be installed. (The default locations for these are volitile, and disappear when logging off.)

The Files pane (usually in the lower right of RStudio) can be pointed at the working directory by clicking on the More dropdown, and clicking on Go To Working Directory

Test the VM

Create some data and a plot

Create a simple data set and a plot. First, create an R script,

  • click on File, then New File and R Script

A new file named Untitled will be created. Save that file by clicking on the disk icon, or by

  • clicking on File, then Save As and
  • enter test01 in the File name: area of the Save File dialog box.

As the file is saved, it will automatically be given the extension .R.

Next, copy or type the following into the test01.R script window.

x <- rnorm(1000)
plot(x)

Save the file again.

Now run the script by selecting the code, and clicking on the Run button (to run both lines at once) or by placing the cursor in the first line and clicking Run, and then clicking Run again as the cursor moves to the second line.

Save the plot as a .pdf

  • click on Export and then Save as PDF...;
  • in the Save Plot as PDF dialog box, the Directory should be R:/geog495_1/Student_Data/userid, where "userid" is again your userid. if the path is not correct, the Directory button can be used to browse to it;
  • click Save

Close RStudio

To quit RStudio,

  • click on File, Quit R Session.

A Quit R Session dialog box will pop up, asking if you wish to save the Worspace image (.RData), and the script file test01.R.

  • click on the Save Selected.

Log off the virtual machine

To log off the virtual machine, use the gray tab at the left of the browser window, and click on the menu button of the Running virtual machine, and click on Logoff. Close the brower to completely disconnect.

Log in, run RStudio, and check that the data have been saved

Start the virtual machine again, and RStudio as above. Copy and paste or type the following where as usual, userid is replace by your userid.

 setwd("R:/geog495_1/Student_Data/userid/") 
 .libPaths("R:/geog495_1/Class_Data/win-library/3.6")

Use Files and More to set the view of the Files pane to the working directory, and you should see the plot file, .Rdata workspace, and the script file, and the variable x should appear in the Environment tab.

An alternative way to start RStudio

Once you have files in your "R" folder (R:/geog495_1/Student_Data/userid/) you can start RStudio simply by clicking on a script file (*.R) or the R workspace file (.RData). A useful thing to do would be to create a script called, e.g. startup.R, and copy into it the setup code:

 setwd("R:/geog495_1/Student_Data/userid/") 
 .libPaths("R:/geog495_1/Class_Data/win-library/3.6")

(replacing userid with your user id). Once that file is created and saved, simply clicking on it will start RStudio, and open that script file. Then the first thing to do is to select the two lines, and run the script. (Note that simply opening the script in RStudio doesn't execute the lines.)