2013/04/10

R is a piece of software, but it is also a community.

Help community

The most visible aspect of the R community is help.  This is also the most useful to new users.  The initial sense of cooperation with R was driven mainly by people helping each other.

You don’t need to actively participate in order to benefit from the help venues — just watching can be very educational.

If you do ask a question, it is in your best interests to formulate the question well.  There are numerous places that give you hints on how to ask a question well, including Circle 9 of The R Inferno and the R posting guide.  Each venue has a slightly different culture, but the main principles are the same.

Stackoverflow

Stackoverflow seems to have become the most popular venue to ask for help with R.  There is an R tag which, by the way, you can subscribe to with RSS.

Warning: Don’t ask questions on Stackoverflow that require opinion.  Stackoverflow is for objective facts.  Go elsewhere (see below) if you are seeking opinions.

Cross Validated

If your question is more statistical than computational, then Cross Validated will be a good place to ask it.

R-help

The R-help mailing list was the original watering hole.  While many people have moved to Stackoverflow, there is still substantial traffic on the list.

Many prefer to look at R-help via a mechanism other than email — gmane for instance.

R-devel

R-devel is a mailing list aimed at the development of R and more technical programming in R.  Some people have the misconception that this list is for a restricted group.  It is open to whomever cares to be a part of it.

R-sig-*

There are several additional mailing lists.  Many of them are sigs, as in “Special Interest Group”.

Package-specific mailing lists

R-forge supports the creation of mailing lists for its projects.  So, for instance, both Rcpp and data.table have their own mailing lists.

Blogs

There are many blogs that talk about R.  You can scour the web for them (there are lots), or you can go to R-bloggers where they are aggregated.  (If you know of a blog that has posts about R but is not on R-bloggers, then please encourage them to join.)

Another aggregator is statsblogs.  As the name suggests it is focused on statistics, but it includes blogs about R.

Twitter

You don’t need a twitter account in order to benefit from it.  You can learn what the twittersphere is saying about R by going to:

https://twitter.com/#rstats

What is happening is that you are doing a search on the hashtag — #rstats — for the R language.  People can make up their own hashtags, #rstats is what has come to be agreed upon for R.

R tweeters

There is a number of people who sometimes to commonly tweet about R.  In good R fashion, they are presented in random order.  That is, via the command:

sample(rtweeters)

@hadleywickham (ggplot2, etc.) — @statalgo — @rOpenSci — @simplystats — @RDataMining — @rbloggers — @statsBlogs — @useR_2013 (the conference in La Mancha) — @inside_R — @cmastication — @rstudioapp — @revodavid — @RLangTip (an R tip a day) — @joshua_ulrich — @pteetor (R Cookbook) — @winston_chang — @romain_francois (Rcpp, R graph gallery, etc.) — @isDotR — @drewconway — @siah — @eddelbuettel (Rcpp, etcl) — @jstatsoft (Journal of Statistical Software) — @pentalibra (R for Dummies) — @Rstats4Econ — @robjhyndman (forecast, etc.) — @StackOverflowR

Oh, and @burnsstat — @portfolioprobe

About Twitter

I started out being quite skeptical of Twitter, but I’ve largely been won over.  While it is true that brevity is not a sufficient condition for wit, you have additional ways to filter out noise.  In particular, you get to choose who you listen to.

The main problem with Twitter in my opinion is that it has the memory span of a goldfish.  A tweet that is an hour old is ancient — even in circles that otherwise have reasonable attention spans.  It’s an interesting phenomenon.  More proof, perhaps, that the medium is the message.

Quora

Quora is a place to ask open-ended questions and questions of opinion.  It includes questions about R.

LinkedIn

LinkedIn has a number of groups devoted to R.  Some of them are:

• The R Project for Statistical Computing
• R Programming
• R/Finance
• R Users subgroup of Statistics and Analytics Consultants
• R Professionals Network

My view of these groups is that they are an excellent answer to an unknown question.  I’m sure that they can be quite valuable, but I’m not sure for what.  Stackoverflow and the various mailing lists are better equipped to handle technical questions.  The groups can be used for the types of questions that are on Quora, but I think there is more potential power in the groups than just that.

About LinkedIn

An error that some people make is to use their work address as their email in LinkedIn.  If you suddenly lose your job, then you won’t have access to your account — and that may well be a time you want it the most.

Package sharing

R does packages well — it is easy to extend the functionality of the language.  RStudio has made it easier than ever before to create packages.

There are three main repositories of packages:

• CRAN
• Bioconductor
• Omegahat

CRAN is the biggest with over 4000 packages.  Bioconductor focuses on genomics data and has over 600 packages.  Omegahat is mainly about connecting R to other languages.

R-forge

R-forge is a nursery for R packages.  Several packages on CRAN also have a development version of the package on R-forge.

Github

Some people put their R packages on Github.  The devtools package (on CRAN) has a function that will install R packages that are on Github:

require(devtools)
install_github("wingui", "halpo")

You need a Github user name to do this.

Google code

Packages are also housed in Google code.

Your own repository

You can create your own repository of R packages for private or public consumption.  You just need to make the directory structure conform to the rules, and add the contents.  There are explanations of what is required.  If you get confused, look at what is in a repository or two.

Conferences

Periodically some people actually meet in person.

useR!

The useR! series of conferences usually toggles between alternate sides of the north Atlantic.  This year it is to be in La Mancha.

R in Finance

There have now been several years of the R/Finance conference in Chicago.

R in Insurance

Coming up is the first R in Insurance conference in London.

Others

R is likely to be talked about at many other conferences.  In particular, the Joint Statistical Meetings and the Royal Statistical Society Conference.

User groups

There are dozens of R user groups around the world.

Questions

What have I missed? (on-topic self-promotion is allowed — even encouraged.)

Epilogue

And the little plastic castle
Is a surprise every time

— from “Little Plastic Castle” by Ani DiFranco