Hacker News Reader

Comment by alister | Hacker News Reader

I might be interested in learning R if I knew what it was. Even the webpage doesn't explain what it is until you watch one of the videos.

I see this all the time on software-related websites. They launch into the technical details without even one sentence about what it is. I could give a whole bunch of examples here but I think everyone has experienced this.

Please include something like the following on the webpage and--if there's space--in the headline too:

"R (a programming language for statistics)"

I don't mean to single out the guy who did these great tutorials. I mean this for everybody. If your software product/OS/service/language is not Photoshop, Linux, Google, or Java, then write a couple of words right at the top to explain what it is please.

View on HN

If this post wasn't on Hacker News, how would someone reach this page without knowing what R is? Would you expect a JavaScript tutorial page to explain what JavaScript is? Or a statistics tutorial page to explain what statistics is?

More broadly, your suggestion makes a lot of sense for websites about software projects. For example, the R software project page (the first hit if you Google for "r") says that R is "a free software environment for statistical computing and graphics." However, when would anyone seek out a set of tutorials about something without knowing (at least at the one sentence summary level) what that thing is in the first place?

Unless you're doing in-depth statistical analysis, there's not really any reason to know anything about R. It isn't a useful general purpose language.

That is to say, while your suggestion is certainly valid and appreciable, the reason why there are rarely primers for content about R (specifically) is that the intended audience is primarily academic, where the audience is already familiar with the technology. if you take an applied statistics class, you'll learn how to use R there.

If you're not already in the club, there isn't much reason for you to join it unless you're leaving industry to do scientific-level statistical analysis.

This isn't me saying "Mind your own business!". I'm just giving the perspective of the writers for most R content.

But couldn't R have applications in business (industry)?

A businessperson once asked me about R after he read a non-technical article about it in a major newspaper. This person has never been in academia. Why was he curious? What could he possibly see in R?

I see R as a "disruptor" to the market that SPSS has had under lock and key for as long as I can remember. Perhaps statistics has utility outside of academics? Maybe R can give anyone the power of something like SPSS, for free.

Now what I want to know is how to convert "screenr" into a more open format. Can you convert to video? R is not dependent on a GUI. Maybe you could convert to screen(1) format, along with an mp3.

Certainly it does. Perhaps I shouldn't have emphasized academics so much.

Anyone who does in-depth statistical analysis can use R. It is equivalent to SPSS, SAS, and matlab in its applicability, plus its free.

People involved in 'big data' and serious analytics in business are already using R a lot of the time. The other group of people in business who use the closed source applications above, however, aren't as technical, and have a pretty big roadblock ahead of them.

The issue is just a question about ease of use. There are companies out there who offer products that compare with the above (http://www.revolutionanalytics.com/), but they still cost money. The 'problem' of R in business is that there aren't any good GUI's for it that allow you to just go menu>ANOVA>3 way factorial>[settings dialog]>Calculate and then get there.

Of course there are a few other reasons why those other applications exist and thrive in spite of R, but I think the limited use cases and lack of a comprehensive GUI are its biggest roadblocks.

Regarding R for BigData, do you think R is going to continue to be a stable of the analytical toolset as data sizes continues to grow?

I use R to prototype machine learning techniques on a small set of data, or visualize some summary statistics. But if I want to run K-Means Clustering or Support Vector Machine algorithms on 1,000,000,000 rows of data, I've found that running R on Hadoop is tricky. There are some libraries out there ( for example, RHadoop https://github.com/RevolutionAnalytics/RHadoop/wiki/rmr ) but they require writing your algorithm in such a manner that algorithms must be adapted to run within map() and reduce() functions. My understanding is that the built-in functions that make R so useful will often not adapt well to a mapreduce algorithm.

From what I've seen, once an algorithm is prototyped in something like R/Matlab, if the data size warrants it, it's best to re-write the algorithm in Java MapReduce or use Apache Mahout.

or C.

RCommander is quite helpful: http://socserv.mcmaster.ca/jfox/Misc/Rcmdr/

I can remember when SPSS had no GUI. It seemed much more responsive back then, running on a VAX.

If we can achieve bigger and better things by teaching a new generation of users how to use the command line again, maybe it's worth considering?

No, absolutely not. Totally unrealistic. Forget I mentioned it.

My favorite thing about R is that is has no GUI. To me it makes it more powerful.

There is a difference between an optional GUI and a mandatory one. I don't see why an optional GUI would hurt R, but I have almost no experience in this area. Would you mind explaining?

I never said an optional GUI would hurt R. (I think optional GUI's are a smart idea.) I was referring to SAS and SPSS. They have mandatory GUI's.

It does. We use it in insurance more and more.

It is not a great general purpose language. I use it to create automated reports and analyses. It's got every statistical test under the sun in the libraries, and the graphs are very flexible, provided you are willing to tweak a lot.

Well between SAS and SPSS they have had the market under lock and key. Yes R is widely used in industry these days. Yes, unless you've got serious statistical work to do it is a fairly horrible language to work with.

The 3 minute tutorial idea is nice. But the problem of the videos are that the author seems to expect already the viewer to have an idea for an applied project.

For someone who just wants to play with it (apparently OP & me), the structure looks extremely un-inviting. I would love to see some project or research problem as early as possible in the process and not entertain myself with juicy videos such as "how to handle NA (missing values) in r: understand and detect NA, remove NA, replace NA with 0" without knowing what when where and why I can apply that.

That's probably my single biggest complaint about Web projects in general. The failure to describe, briefly and clearly, what a tool does.

Not a problem in this case (I've dabbled with R for over a decade), but I definitely feel your pain and agree with your criticism.

The HN default of using a vague headline and often linking directly to a site with little or no explanatory context (the fact that you can either post a link, or text, but not both) really doesn't help this.

Reply on news.ycombinator.com