Lab 13
Programming for Data Analysis
Lehman College, City University of New York
Spring 2017

Today's lab focuses on standard statistical tools for comparing data series and useful packages for plotting in the R programming language.

Using Packages

As with Python, R has both built-in functionality as well as a large community writing additional, useful packages. R packages can be found at the Comprehensive R Archive Network (CRAN) repository.

In lecture, we used both the built-in (base) graphics package and a fancier graphics package to make prettier images, ggplot2. Here's directions on how to load it onto your computer:

• At the R console (this is different from adding packages to Python, where we were at the terminal window-- for R, you are running R when downloading new libraries).
• The command to install packages is: install.packages(). For example, to install ggplot2, at the R console, you type:

  	install.packages("ggplot2")
  	

• You will be asked to choose a mirror site from which to download the package. The closer the mirror is to your location, the faster the download (for example, choosing London, UK will take longer than Indiana, USA, since the files would have to transferred across the ocean instead of using (relatively) local routing through the US).
• To use a package, type at the prompt (or include in your file):

  	library("ggplot2")
  

Today's lab uses ggplot as well as GGally to make visualizations of correlations in a dataframe. Repeat the steps above to download and install GGally.

Built-in Datasets

R has a package of built in data that can be used for examples and analysis at R's Dataset Package.

Some packages that we have seen (or will see) in lecture, lab, and programming include:

• cars: Speed and stopping distance of cars, recorded in 1920's
• mtcars: 1974 Motor Trend magazine's car analysis.
• presidents: US presidential approval ratings from 1945 to 1974.
• Titantic: A manifest of the Titanic and whose survived.
• quakes: Locations of a 1000 earthquakes off the island of Fiji.
• volcano: Digitized topographic map of Mt Eden (Maunga Whau) near Auckland, New Zealand.

ggplot2

In lecture, we briefly touched on the Harvard's R graphics tutorial. Work through the tutorial, and then try the related challenges/programming problems.

Comparing Data Series

When analyzing data, a common question is do two things have intertwined behavior? That is, if one is larger than average, would you expect the larger than average? There are rigorous tools to measure this inter-related behavior. We will focus on an incredibly useful one that is easily used in R, called correlation. Work through the DataCamp's correlation tutorial. Then try the related challenges/programming problems.

In-class Quiz

During lab, there is a paper quiz on CW #25: Programs 41-45

What's Next?

If you finish the lab early, now is a great time to get a head start on the programming problems due early next week. There's instructors to help you and you already have Python up and running. The Programming Problem List has problem descriptions, suggested reading, and due dates next to each problem.