Problem Sets

Return Home

Contact Us

Steven Holland

Problem Sets

Due date: All problem sets are due at 2:00 PM on the date specified for each assignment, unless stated otherwise. The timestamp on your email, as displayed in Outlook on my computer will be used to determine whether problem sets were submitted on time. I recommend completing these problem sets early and not risk waiting until the due date to finish them.

What to turn in: Save your R commands as a plain-text (.txt) file named xxxx-2.txt, replacing xxxx with your last name, lowercase. Replace the number with the number of the problem set. For example, if your name is Rita Colwell and you are turning in problem set #4, your file should be named colwell-4.txt.

Remove unnecessary commands from your file. For example, if you try several commands that do not work or that produce errors, delete them. If you are testing code that is not needed to complete the problem set, delete that too. Your commands file should contain all the commands and comments needed for the problem set, and no more. Get in the habit of cleaning up your own code the same way.

Follow the advice on Becoming a better coder. Based on comments on that homework, I will update that page after most problem sets are returned. Once something is posted to this page, you must incorporate those concepts in your code. Not doing so will be treated as a mistake and points will be deducted from your homework. Over time, the quality of your code will improve substantially by following these instructions.

How to submit your problem set: Email your problem set to stratum@uga.edu, and the subject line to 8370 problem set XXXX, replacing XXXX with the problem set number (e.g., 8370 problem set 1). If you do this correctly (address is correct, subject line is correct, an attachment is included), you will receive an automated response saying that your problem set was received.

Please submit your problem set only once. I realize that sometimes you will discover errors that you want to correct. That is fine, but do not make a habit of resubmitting your work. I will always grade the last version submitted before the deadline, and I will not look at earlier submissions or resubmissions after the deadline.

Grading: Each problem set is worth 20 points.

To grade your problem sets, I will copy your commands into an R session and run your code. Your code must run without errors. You must verify that the commands in your commands file do not generate errors. The best way to do this is to close R, restart R, and paste all your commands into the R window. If doing this produces errors, fix your commands file and repeat this procedure until it works flawlessly. This may sound cumbersome, but it is a habit we will build so that when you share code with colleagues, it always runs without errors.

Embedding paths and using setwd() is guaranteed to generate errors on my computer, even though it will run flawlessly on yours, so never embed paths in your code. Similarly, device-specific ways of making plotting windows may also generate errors on my computer while running fine on yours. Always use dev.new() instead of specific methods for macOS, Windows, and Linux.

Points will be deducted as follows:

• Late work: -10 points (half the grade) per 24-hour period after the due date and time. For example, if the work is due at 6:00 PM, 6:01 is late.
• Including a path in your code: -3 per occurrence
• Incorrectly submitted file (wrong e-mail subject line, wrong file name, submitting a file that is not a plain-text .txt file, etc.): -1 point per error
• Code that generates errors or warnings: -2 points per class of error. If your code generates an excessive amount of errors, it will be returned for resubmission and receive the penalty for late work.
• Including unnecessary commands: -1 point, increasing with the number of unnecessary commands
• Failing to call detach() after you call attach() on a data frame: -1 point
• Mistakes: -1 point each

Asking Questions

When you cannot solve a problem and need help, a few things can make it easier to get an answer quickly.

First, state the problem succinctly, including (1) what you did, (2) what you expected to happen, and (3) what happened:

• For what you did, give the command exactly as you typed it, as well as any necessary commands before that; delete the commands that are not needed. If a data file is required, send a copy of that data file if I do not have it.
• For what you expected to happen, state simply what you thought the command should have done.
• For what happened, give the R output, including all of the text of any error message. Copy and paste the text from your R session, instead of taking a screenshot. The text version makes it easier for me to replicate your work exactly.

I will usually try to replicate your error. If you don't supply your command, necessary data files, and the exact error message, I have no way of knowing whether I’m getting the same result as you. If you supply all of these things, I will generally be able to answer your question in a single email.

Finally, remember to start the subject line of your email with 8370.