[Solved] STAT480 Group Project

The group project is based on the airlines data set from the US Department of Transportations Bureau of Transportation Statistics (BTS). This is the data we looked at in Chapter 5 of Data Science in R. To get the data for certain years, we will use the same script file for downloading particular years.

The airlinesauxiliaryfiles.zip file in the compass course space includes a ReadMe file describing the full data set, details on how to download particular years of data, and descriptions of the auxiliary files included in the zip file. The auxiliary files include airport codes, carrier information and plane information which might be useful to combine with the raw data.

Each group will be assigned two years of data to analyze. The assigned years are as follows:

The final project will count for 30% of your final grade.

Each group will be working with a little over 1 GB of data. Processing and analysis of the data must make use of multiple software applications (including Hadoop-based applications) we have used in the virtual machine.

The general goal is to extract interesting information, trends, and comparisons about flights in the years your group is assigned. You may need to deal with missing values in variables and perhaps variables that are entirely missing.

Some possibilities you might consider for your two years:

• Cancellation trends
• Trends in delayed arrival or departure
• Trends in number of flights
• Other trends you think might be interesting

You should consider these types of analysis and any others you find interesting to give a broad overview and comparison of travel in your two years. Analysis should focus on comparing between the two years and with respect to other classifications. For instance, you might consider trend or overall analysis by region, airport, carrier, plane manufacturer or anything else you think might be interesting. Models as a function of categorical factors (region, carrier, etc.) might also be useful for exploration and analysis.

Interesting features in time should be explored for historical significance. For instance, if there is a

noticeable difference in cancellation rates or delays on a particular day or week, was there a major storm or other historical event that might be related?

Interesting visualizations must also be provided. This should include visualization beyond the usual scatter plots and bar charts and might include heat maps, tree maps, network visualizations (e.g. to see connectivity and/or identify carrier hubs), or other advanced visualizations. Some of these types of visualizations will be discussed in class, but you are not limited to visualizations discussed in class.