Customer Reviews

Interactive and Dynamic Graphics for Data Analysis: With R and GGobi (Use R!)

5 star:		(0)
4 star:		(1)
3 star:		(1)
2 star:		(0)
1 star:		(0)

 

 

This is definitely not an introductory tutorial or a how-to book on data visualization. The authors are the developers of the wonderful GGobi visualization system and its R integration and they have put lots of documentation online to which they don't hesitate to refer in many parts of the book. That means you have to have some background in R programming, statistical terminology, principal component analysis, etc.

The main strength of the book is providing very good examples depicting how dynamic graphics based analyses can help analytical methods. I especially consider the chapters on supervised classification and clustering very well designed. Many critical aspects are stressed and the importance of "looking at data before diving into support vector machines, linear discrminant analysis, decision trees and self-organizing maps" is shown from different perspectives (pun intended ;-)

If you are serious about data visualization, data mining and statistics then this book *along* with the accompanying website will be a very good guide. The exercises at the end of each chapter will also provide challenges as well as valuable insights.

This book helped me get into ggobi, which is what I bought it for. Two main obstacles for new ggobi users are its peculiar user interface and awkward input of data. The latter issue is solved by the R package rggobi, which can use R data frames directly. The user interface is marginally better with rggobi; you can script the basic creation of figures, selection of datapoints, and choice of focal variables. The whole thing is still rather awkward, though: For example, to change views or interaction modes, you need to leave the figure window and visit the main ggobi window. While it is possible to change formatting (glyphs, colours, sizes) it is very difficult to figure out. The book is almost indispensable in taking you through the rather unintuitive workflow of (r)ggobi. The book also has OK introductions on classification and clustering. All in all it served its purpose, but I probably won't be coming back to it.

Users looking for an alternative might try the R package iplots, which has a somewhat smoother user interface; I ended up using both.