Borasky Research Journal Google+ Page Data Journalism Developer Studio 2012LX Blog
Disclosure
As you probably know, I live in the Portland, Oregon area and have for many years. One of the must-visit places here is Powell’s Books. The book links in this post will all take you to Powell’s as part of their Partner Program. If you’d like to join the program too, here’s the link.
Updated September 11, 2011: I recently purchased a Kindle, and two of these books are now available in that format. For those two, I’ve tweeted my Amazon Affiliate links out and have embedded those tweets here.
Prerequisite Software
To get the most out of these books, you will need to install some software. You will need Mondrian, R, GGobi, and the ggplot2, rggobi and DescribeDisplay R packages. All of these will run on a Windows, Macintosh or Linux desktop / laptop, including most netbooks. And they are all free, open source software. An easy way to get them all, packaged in an openSUSE Linux appliance, is to download Data Journalism Developer Studio 2012.
Ggplot2: Elegant Graphics for Data Analysis (Use R)
ggplot2 is an advanced graphics package for the R programming language. It is based on the grammar of graphics (Grammar of Graphics 2ND Edition). ggplot2 generates the most beautiful static graphics I’ve ever seen. You can use ggplot2 at any stage of your analysis. Simple exploratory plots can be made with a simple call to the “qplot” function, and when you’re ready to create a final report or presentation, you can get publication-quality graphics.
The two things I like most about the ggplot2 package are
- The absolutely stunning visual appeal of the plots it produces: Dr. Wickham has paid great attention to the visual aspects of the output. I don’t know of another package in any language that generates such beautiful plots.
- The numerous built-in analysis methods: Boxplots, kernel and quantile regression and smoothing, faceted plots – all are “standard equipment” with ggplot2.
Interactive Graphics for Data Analysis: Principles and Examples
This book is a complete course in interactive graphics for data analysis. It is mostly based on the Mondrian interactive statistical data visualization system, although there is some use of R as well. The first part covers the basic tools, and the second part gives case studies.
The case studies really are the best part of the book. They cover geographical analysis, some interesting history from the sinking of the Titanic and the 2004 Florida election. As I note below, there is some overlap in tools between Mondrian and GGobi, but you really need both books and both packages to be able to do everything.
Interactive and Dynamic Graphics for Data Analysis: With R and Ggobi (Use R)
As the title implies, this book is also a complete course in data analysis using interactive graphics. But the focus here is on R and GGobi rather than Mondrian. While there is some overlap in the tools, there are some things Mondrian does that GGobi doesn’t do, and vice versa. A partial list:
- Geographic datasets: Mondrian only
- Mosaic plots: Mondrian only
- Classification: GGobi only
- Clustering: GGobi only
- Social network graphs: GGobi only
In addition, GGobi integrates directly with R and ggplot2 via the rggobi and DescribeDisplay packages. There are some integration points between R and Mondrian, but that integration isn’t as tight as it is with R, GGobi and ggplot2.
