afterword Into the rabbit hole

We’ve covered a broad range of topics in the book, including major ones like the R development environment, data management, traditional statistical models, and statistical graphics. We’ve also examined hidden gems like resampling statistics, missing values imputation, and interactive graphics. The great (or perhaps infuriating) thing about R is that there’s always more to learn.

R is a large, robust, and evolving statistical platform and programming language. With so many new packages, frequent updates, and new directions, how can you stay current? Happily, many websites support this active community and provide coverage of platform and package changes, new methodologies, and a wealth of tutorials. I’ve listed some of my favorite sites next:

■The R Project (www.r-project.org)

The official R website and your first stop for all things R. The site includes extensive documentation, including “An Introduction to R,” “The R Language Definition,” “Writing R Extensions,” “R Data Import/Export,” “R Installation and Administration,” and “The R FAQ.”

■The R Journal (http://journal.r-project.org)

A freely accessible, refereed journal containing articles on the R project and contributed packages.

■R Bloggers (www.r-bloggers.com)

A central hub (blog aggregator) that collects content from bloggers writing about R. It contains new articles daily. I’m addicted to it.

■Planet R (http://planetr.stderr.org)

Another good site aggregator, including information from a wide range of sources. Updated daily.

532

AFTERWORD Into the rabbit hole

533

■CRANberries (http://dirk.eddelbuettel.com/cranberries)

A site that aggregates information about new and updated packages and contains links to CRAN for each.

■Journal of Statistical Software (www.jstatsoft.org)

A freely accessible, refereed journal containing articles, book reviews, and code snippets on statistical computing. Contains frequent articles about R.

■Revolutions (http://blog.revolution-computing.com)

A popular, well-organized blog dedicated to news and information about R.

■CRAN Task Views (http://cran.r-project.org/web/views)

Task views are guides to the use of R in different academic and research fields. They include a description of the packages and methods available for a given area. Currently, 33 task views are available (see the following table).

CRAN task views

Bayesian Inference	Natural Language Processing
Chemometrics and Computational Physics	Numerical Mathematics
Clinical Trial Design, Monitoring, and Analysis	Official Statistics & Survey Methodology
Cluster Analysis & Finite Mixture Models	Optimization and Mathematical Programming
Differential Equations	Analysis of Pharmacokinetic Data
Probability Distributions	Phylogenetics, Especially Comparative Methods
Computational Econometrics	Psychometric Models and Methods
Analysis of Ecological and Environmental Data	Reproducible Research
Design of Experiments (DoE) & Analysis of Experi-	Robust Statistical Methods
mental Data
Empirical Finance	Statistics for the Social Sciences
Statistical Genetics	Analysis of Spatial Data
Graphic Displays & Dynamic Graphics & Graphic	Handling and Analyzing Spatio-Temporal Data
Devices & Visualization
High-Performance and Parallel Computing with R	Survival Analysis
Machine Learning & Statistical Learning	Time Series Analysis
Medical Image Analysis	Web Technologies and Services
Meta-Analysis	gRaphical Models in R
Multivariate Statistics

534	AFTERWORD Into the rabbit hole

■R-Help main R mailing list (https://stat.ethz.ch/mailman/listinfo/r-help)

This electronic mailing list is the best place to ask questions about R. The archives are also searchable. Be sure to read the FAQ before posting questions.

■Cross Validated (http://stats.stackexchange.com)

A question and answer site for people interested in statistics and data science. This is a good place to post questions about R and see what other people are asking.

■Quick-R (www.statmethods.net)

This is my R website. It’s stocked with more than 80 brief tutorials on R topics. False modesty forbids me from saying more.

The R community is a helpful, vibrant, and exciting lot. Welcome to Wonderland.

appendix A Graphical user interfaces

You turned here first, didn’t you? By default, R provides a simple command-line interface (CLI). The user enters statements at a command-line prompt (> by default), and each command is executed one at a time. For many data analysts, the CLI is one of R’s most significant limitations.

There have been a number of attempts to create more graphical interfaces, ranging from code editors that interact with R (such as RStudio), to GUIs for specific functions or packages (such as BiplotGUI), to full-blown GUIs that allow you to construct analyses through interactions with menus and dialog boxes (such as R Commander).

Several of the more useful code editors are listed in table A.1.

Table A.1 Integrated development environments and syntax editors

Name	URL
RStudio	www.rstudio.com/products/RStudio
Eclipse with StatET plug-in	www.eclipse.org and www.walware.de/goto/statet
Architect	www.openanalytics.eu/architect
ESS (Emacs Speaks Statistics)	http://ess.r-project.org
JGR	http://jgr.markushelbig.org/JGR.html
Tinn-R (Windows only)	http://nbcgib.uesc.br/lec/software/editores/tinn-r/en
Notepad++ with NppToR (Windows only)	http://notepad-plus-plus.org and
	http://sourceforge.net/projects/npptor

These code editors let you edit and execute R code and include syntax highlighting, statement completion, object exploration, project organization, and online help. A screenshot of RStudio is shown in figure A.1.

535

536	APPENDIX A Graphical user interfaces

Figure A.1 RStudio IDE

Several promising, full-blown GUIs for R are listed in table A.2. The GUIs available for R are less comprehensive and mature than those offered by SAS or IBM SPSS, but they’re developing rapidly.

Table A.2 Comprehensive GUIs for R

Name	URL
JGR/Deducer	www.deducer.org
R Analytic Flow	www.ef-prime.com/products/ranalyticflow_en
Rattle (for data mining)	http://rattle.togaware.com
R Commander	http://socserv.mcmaster.ca/jfox/Misc/Rcmdr
Rkward	http://rkward.sourceforge.net

My favorite GUI for introductory statistics courses is R Commander (shown in figure A.2). Finally, a number of applications allow you to create GUI wrappers for R functions (including user-written functions). These include the R GUI Generator (RGG)

APPENDIX A Graphical user interfaces

537

Figure A.2 R Commander GUI

(http://rgg.r-forge.r-project.org) and the fgui and twiddler packages available from CRAN. The most comprehensive approach is currently Shiny (http://shiny.rstudio

.com/). Shiny lets you easily create web applications with interactive access to R functions.