The R Language The Good The Bad & The Ugly • John Cook • GOTO 2012

This presentation was recorded at GOTO Aarhus 2012. #gotocon #gotoaar

John Cook - Research Statistician at M. D. Anderson Cancer Center

ABSTRACT
R is a domain-specific language for analyzing data. Why does data analysis need its own DSL? What does R do well and what does it do poorly? How can developers take advantage of R's strengths and mitigate its weaknesses? This talk will give some answers to these questions.

#Rlang #Rlanguage #DomainSpecific

Looking for a unique learning experience?

SUBSCRIBE TO OUR CHANNEL - new videos posted almost daily.

[CradleToGraveR]

I love the analogies at the begging. Learning SQL without wanting to learn about databases. Ha ha. Great lecture. I learned academic level C++ first and years later picked up some R. You really showed me a different perspective when it comes why R is so awesome.

[geraldcreedon]

Excellent introduction / description of the advantages and pitfalls of using R.

[avroB]

When discussing the efficiency of programming languages, especially those for use as experimental tools, human time is multiple orders of magnitude more important than machine time. Bit-nanoseconds cost infinitesimal amounts, human time is not infrequently expensive in itself, and subject to multiplication by the cost of project delay.

Worrying about computer "efficiency" is mostly a legacy of 1960s thinking, when computers were expensive and people cheap.

[jmhjcfk]

You could avoid the waste of memory by typing : x<-sum(rnorm(1e6)) at 26:18

[jacob]

"One may get headaches and even gag the fist few times" I never got that, but I love R.

[russkee]

2012, this is old. The latest R language is super good with all the libraries

[danielgallant]

The guys has something wrong... act check: many scientists have been using R for statistics work, not just statistics research, and we have been doing so for years. This is reflected by the large quantity of statistical packages developed. these are developed for us the researchers. The flexibility of R was our solution to the limited options in end-user software. As an ecologist I have been using R since 2006 or so.

[OleksandrFialko]

"The alternative to R isn't Python". Well, Python is alternative to everything.

[dreznik]

would love to see an updated version (late 2018) of this talk, though my feeling is most of what has been said is still true.

[morenoaf]

Really good and useful presentation, it is a very honest and professional review of the tool

Thanks,

[quartzimaging]

Dude starts off saying people use R for statistical research but not for analysis. Three minutes later he shows data that says R is the by far most popular analysis tool. wtf?

[avroB]

Any programming language, no matter how specialised or domain-specific, will be used to write a least one application for which it is utterly unsuitable.

Sometimes, this is because the author doesn't know any better. Other times, the motivation is sheer programmer perversity.

[mikejmc]

This is a very nice overview of R for the beginner. I won't pick apart the language "gaffes" when the speaker defaults to more common programming-speak (e.g., use of term "modules"), as an audience familiar with other languages is more likely to understand those terms as opposed to the quirky R terminology, and yes, explicit loops in ANY interpreted language usually are the devil. However, as someone who has programmed in R and SAS for 10+ years now, I do take issue with the characterization of SAS as being on par with Windows Command Prompt in terms of data management functionality. That's extremely inaccurate, but I'll give him a pass on that one, as he admits to having no prior experience with SAS.

[akshaysingh]

Oh my god we are missing data..
what are we gonna do throw an exception call the police...:)

[dwinsemius]

This guy doesn't know R very well. Assignment of a data-object to a function name does not mask the function. For-loops are not inherently inefficient (this was fixed about twenty versions ago). You should not use attach(). R uses the term "package", not "module". The "parallel" package was built into the base distribution several versions ago (which does use parts of R). He was way behind the available books at the time this talk was given. And... how can you give a talk on R and mangle Prof Brian Ripley's name?

[WahranRai]

The simplicity and understanding of R were destroyed by H. Wickham with his notions of pipe %>%, tidyverse etc..

[cepi]

can you recommend good books to learning statistics with R?

[almanac]

Did anyone else catch the s-- joke at around @3:22

[surjerrylee]

Do you need a background in statistics to use R?

[ahmgeek]

The R Language The Good The Bad And The Ugly