- 4/15/2014 5:10:14 PM
Ah, thanks for hunting that down Pierre. I haven't tuned in yet, but I did take a look at a "Related story" that I found raises quite a few good points: 5 Key Considerations When Choosing Open Source Statistics Software. As the writer notes, "just because something is popular, or free, doesn't mean that it's right for your needs." What follows is a good "rough guide to some of the core issues impacting serious users, wanting to perform serious statistical analyzes in a commercial environment." Great points to really think about for those considering open source and commercial alternatives.
- 4/15/2014 4:56:24 PM
Found it - I had the sites confused. Data Informed was the site with the article. The article notes a Rexer Analytics survey in which the average data miner uses 5 different tools, and R was among 70 percent of the responses.
- by bkbeverly, Data Doctor
- 4/14/2014 3:19:14 PM
This thread reminded me that in pharmaceuticals, the drugs that you receive from your doctor depends on which set of health practice laws or pharmaceutical reps influence the doctor's choice. In like manner, I could see the analytics choice of weapons dependent upon who can best influence the dean or the functional manager. Historically, the sales reps started with the developers until they discovered that they did not have authority to make purchases. So then the sales reps started contacting the managers to sway software sales. The young grad students are showing that the university deans are the new points of influence. The deans control what the university will use as the basis of professional training. What is interesting is that the deans are higher up in the food chain - one's academic environment shapes one's analytic orientation long before one enters the job market. Hence if the schools make an undocumented decision (based on the academic discipline of course) to endorse a specific analytic approach or a buffet of approaches, then the job market will either have to shift to the skill sets available or use the Jedi mind trick to convince the graduates the company way is the right way. In any case. the functional manager is no longer the only person of influence. The game is changing. Whoever controls the schools can also influence the tools.
- by urbie4, Data Doctor
- 4/14/2014 2:08:58 PM
These days, when I'm looking to learn anything software-related, I just go to YouTube and search, e.g.: R TUTORIAL, and you find tons and tons of stuff. A lot of them are not that well done, lousy presentation, etc., but just keep trying them until you find one you like. I use that for almost any application -- if I'm stuck in Final Cut Pro X, or Hype, or Word, or whatever it is, there's always someone who has free tutorials about it. Scary, if you're in the e-learning business!
My expertise with R is at the beginner level, and I'm not much of a programmer in any language -- but R is not hard to learn, at least in terms of doing basic Stats 101 type stuff.
- 4/14/2014 8:52:16 AM
Hi Pierre, don't the benefits of increased processing power, storage capacity, and portability hold true for most any programming language that's been around for a while?
- 4/14/2014 5:45:19 AM
To Bryan's point, many of the young scientists can use R because of the advances on computer storage and processing power. Computers can process much faster than when R was first introduced, and the language is portable. Plus there are larger hard drives, as well as cloud storage, to permit larger data sets to be examined. These are recent developments compared to R's history, which is now about 40 years old, via Roger Peng at John Hopkins, who developed a great introduction course for the university's public health program. This makes R have a new appeal that was not possible at the beginning.
- 4/14/2014 4:22:04 AM
Tomsg, That is a good point about open source in this instance - the idea that its fucntionality is dependent on what users have developed or what you can develop on your own. R is very dependent on user results on its function packages.
- 4/12/2014 7:07:47 AM
It may more about avoiding the cost of adopting a new language. I see businesses that try to extend their tech investment as if it were an asset to be "consumed" - that can be understandable for companies using SAS despite R's "no-cost" open source quality because of the cost of finding and training people combined with the developmental nature around the language (ie its value is based on the programs developed). Plus I recall reading a post on R-bloggers that noted that more data scientists are using R wih another program, suggesting that R has been supplimental. Will look for it and share.