I will focus on the Airline Delays data described on the Data page of the Workshop website and more fully here http://stat-computing.org/dataexpo/2009/the-data.html. The data provide details including delay times for all domestic flights in the USA.

I have been to working with Oakland and San Francisco airports. It would be most interesting for students to use local airports so some description of getting the data for different airports would be nice to provide (whether this step is carried out by students would depend on the level of the class).

A guided assignment might provide specific questions to be answered using simple subsetting, summarizing and visualization. Some examples:

• Which airport has the most delays? (Note that there are different ways of defining what constitutes a delayed flight). Which airport has the longest delays? Does this change over time?
• Are there any origin/destination combinations that are unusually bad for delays?
• Are some days of the week or times of the day particularly bad for delays? Is this different at different airports?
• Is there correlation between the delays at different airports?

A project might be more open ended to encourage students to both ask and answer questions visually (perhaps after some of the previous questions being worked through in class/lab).

A slightly more involved project could explore the relationship between weather conditions and flight delays. The site, http://www.wunderground.com/history/, provides historical weather reports for airports in a nice format for reading into R (almost csv). Getting this data from the web into R would expose students to the functions paste, url, and read.csv and an opportunity to write then apply a function over a range of arguments. This could then be followed by visual exploration to answer predefined or student generated questions.