Learn R for Free

It never hurts to increase your skill set, especially in this economy. And when you can do that for free, so much the better. And when you can learn something for free that is in demand all over Wall Street, well that's just the icing on the cake. So I wanted to let you know about a new course that
Codeschool is offering called Try R as an introduction to the R programming language.

For those who don't know, R is quickly gaining steam among the quant trading set on Wall Street and was recently listed among top skills on Wall Street's wish list. The way it's been explained to me is that R isn't really Excel on steroids, it's more like Excel got exposed to a near-lethal dose of gamma radiation and morphed into the Hulk. So if you have a head for coding, it's definitely something you'll want to check out.

One of the cool things about Codeschool (and several other training sites, to be fair) is their in-browser programming environment. So you don't need to download anything and clog up your hard drive with DE's and a bunch of other junk. In fact, you do all your programming right there on your screen within the course. Now that I think about it, Khan Academy offers the same set-up in their new Computer Programming track.

I realize this won't appeal to everyone (some of us aren't as quant-inclined as others), but it might even be worth checking out the course to gain a little mile-high overview of R. It doesn't hurt to be able to talk about it even if you're no expert.

Anyway, let me know if that helps anyone out here. I try to keep my ear to the ground for things that can make you more competitive or make your lives easier, and maybe this is one of those things. If not, it didn't cost you anything.

For the record, I've taken courses at Codeschool before and they're very well done.

Good luck!

 

@Addinator I think a couple of the other ones are free as well. I want to say the intro to Rails course is free. Anyway, Codeschool is $25 a month for unlimited access to all their courses.

@surferdude No, this isn't something to put on your resume, not unless you get really good at it. This is just an intro course to brush the broad strokes. That said, spending your free time learning R is pretty smart if you know you ultimately want to work in trading. The problem with listing a course like this on a resume is that you run the risk of coming up against an interviewer who really knows his shit and then you're screwed.

 
Best Response

Did this class last night (I already know R; I just wanted to see how low the barriers to entry to learning it had become). It's actually a pretty solid, quick overview. Obviously, you'd need to complement this with a book, and there's nothing like a genuine need (i.e., a project at work, a statistical relationship you're trying to get at, etc.) to drive the points home. Still, there are worse ways to kill time.

To echo Eddie's points: Get technical -- even if you're in IBD and the most complicated thing you do is figure out if the decrease in accounts payable increases cash or decreases cash. If you move onto a hedge fund role or -- hell, even a corporate strategy role these days -- and you're not prepared for the Big Data Revolution the Tech kids are talking about, then at some point, someone who did choose to get technical is going to eat your lunch. We need more multi-faceted people in this world.

Luckily (or unluckily, depending on your bent), I don't really believe Finance is a field that would benefit from turning into a quantitative pseudo-science. (We've all seen where that's gotten Economics). There's always -- always -- going to be a need for people who do the fundamental analysis well, and all the R in the world isn't really useful if you don't have an underlying passion for/understanding of the business environment in which you're operating. The world is also much too complicated to be modeled using least squares regression, though I'm sure there's a Data Miner somewhere trying to prove me wrong, with a model infinitely more complicated. Still, choosing to ignore these recent developments is like a baseball scout choosing to ignore Sabermetrics in a post-Moneyball world. Do so at your own peril.

 
Edmundo Braverman:
^^^Great points. Quick question: having gone through the course already, where would you say someone starting fresh ends up by the end of the course? I mean, what level of understanding/proficiency does the course provide?

I'm probably going to run into a bit of Professor syndrome -- i.e., I know it all so it's hard to see it from the perspective of someone just starting out. However, what I can safely say is that things that took me hours and hours to get down, this class at least lets you know exists, and how to utilize them simply, so that you can approach them with some level of understanding, and probably target your search more effectively. (Two quick parenthetical comments: (1) Programming is basically getting really good at how to find things on StackOverflow; don't let anyone tell you differently; (2) I had to teach myself R so that I could take the upper division Statistics courses at my school; R wasn't a standalone class and they just expected you to learn or be left behind. Not surprisingly, I went to a public school, and I spent a lot of time on Google).

For example, reading files into R, for someone who has never programmed before, is actually a bit of a pain to figure out. Finding a self-contained tutorial online is hard, so you'll usually do something like... 'How do I read a file into R?'... 'Okay, cool, I have this thing. How can I use this thing.' And it's just a lot slower and more, as Nassim Taleb would put it, 'stochastic tinkering.' This class alleviates the pain.

I wish they'd covered the 'apply' family of functions since that is a crucial element of R, and something that makes it relatively different from other programming languages.

Still, hard to complain when something is so brief, so in demand and so free.

TL;DR: Lets you know which things exist and how to use them. Also look up apply functions. Wish MOOC's had existed earlier.

 

What a coincidence. I began learning R very recently because I heard it was used by quants and simply wanted to try out something new. Would you say it's more useful to know than MATLAB? I've been learning both, but I'd like to focus on one.

Thanks for the post

 
Pancakes:
What a coincidence. I began learning R very recently because I heard it was used by quants and simply wanted to try out something new. Would you say it's more useful to know than MATLAB? I've been learning both, but I'd like to focus on one.

Thanks for the post

Isn't there some cost associated with using MATLAB? R is 100% open source and free. I seem to remember something about MATLAB costing something.

Just FYI: I have nothing to do with any of this stuff. It's all way over my head. So if you're wondering if I'm taking this R course, I'm not. Fortunately I've attained an age and level of success that I can pay somebody else to learn it if I need it.

 

R and Excel aren't really that comparable, given the usage and amount of data each can hold (and still operate at a reasonable speed). I've done multiple regressions in R with well over 7000 rows and 50 columns, and it has held up gracefully. Excel on the other hand stumbles over itself trying to open.

R is more on the level of your typical statistical software packages -- Stata, SAS, the works. All three are better to know than MATLAB - I can't think of anyone in Economics or Finance who uses MATLAB over SAS or Stata. R itself has quite a steep learning curve, in my opinion, but once you grasp it, it's amazing to use. I can't see most people in finance having a need for it besides the quants, though.

Currently: future neurologist, current psychotherapist Previously: investor relations (top consulting firm), M&A consulting (Big 4), M&A banking (MM)
 
chicandtoughness:
R and Excel aren't really that comparable, given the usage and amount of data each can hold (and still operate at a reasonable speed). I've done multiple regressions in R with well over 7000 rows and 50 columns, and it has held up gracefully. Excel on the other hand stumbles over itself trying to open.

R is more on the level of your typical statistical software packages -- Stata, SAS, the works. All three are better to know than MATLAB - I can't think of anyone in Economics or Finance who uses MATLAB over SAS or Stata. R itself has quite a steep learning curve, in my opinion, but once you grasp it, it's amazing to use. I can't see most people in finance having a need for it besides the quants, though.

Agreed, never heard of a non-quant using it. except maybe industry Business Intelligence/Analyst roles.

 

The full version of MATLAB costs a small fortune--about $2,000 I believe. You can get a free trial though, and there's a relatively inexpensive student version available for $99.

 

Oh, and I would also add that this class teaches you how to be a 'user' of R, and not so much an 'R programmer.' A subtle point, but one worth mentioning. Luckily, for most people, that's probably going to be enough unless they become more interested/need to do much more complicated things (in which case you're probably going to need to learn other programming languages as well).

 

R is very very nice, but honestly slow. We started running into problems where our models (which had decently efficient algorithms) were taking about 5 minutes to readjust, given no big changes to the underlying data set. If that happened, we had to rebuild which took 20 minutes. The big advantage of R is since it's a statistical program at heart, there's tons of great packages people have added on to cover some things you'd really like (parallel processing and some polynomial stuff were the big things).

Learning that and something like Python/C++ if you ever want to do complicated stuff with big data would be very good (aka not IBD level models). And, Python means you won't hate yourself. Because I really hate myself when I try to learn C++. I think since Python is very nice, people are developing modules to support some important features for it. We saw a parallel processing module that was very nice, but even without using it, Python was faster than R using parallel processing (again, that's a scale thing though. As our models got bigger, I'm sure we would have had to to parallel in python too).

For those of you who know some R stuff, check out a package called pandas for Python. It adds data tables which are similar to R, but work SO MUCH faster (for example, the melt function is much faster. It literally shaved minutes off our computations).

 
canas15:
For those of you who know some R stuff, check out a package called pandas for Python. It adds data tables which are similar to R, but work SO MUCH faster (for example, the melt function is much faster. It literally shaved minutes off our computations).

Not sure, tried doing some analysis using Pandas + Python and the damn thing couldn't even read the file due to the size. R took about a minute to read the file and about 24 hours to conduct the tests. Mind you, it was a 1.X TB file containing currency tick data. The packages in R run circles around python.

If anyone is using R, make sure you load up R Studio.

You can also try Octave, which is an opensource replica of MATLAB: http://www.gnu.org/software/octave/

 
Macro Arbitrage][quote=canas15:
For those of you who know some R stuff, check out a package called pandas for Python. It adds data tables which are similar to R, but work SO MUCH faster (for example, the melt function is much faster. It literally shaved minutes off our computations).

Not sure, tried doing some analysis using Pandas + Python and the damn thing couldn't even read the file due to the size. R took about a minute to read the file and about 24 hours to conduct the tests. Mind you, it was a 1.X TB file containing currency tick data. The packages in R run circles around python.

If anyone is using R, make sure you load up R Studio.

You can also try Octave, which is an opensource replica of MATLAB: http://www.gnu.org/software/octave/[/quote]

Wow. To be fair, the files I used were on the order of ~100mb. Good to know R is still better for really large scale data sets.

 

I used STATA and Matlab extensively in college for regressions in economics... Very similar to R, and all are wonderful tools for statistical analysis. Great for anyone quantitatively focused or just wants to learn some small coding

And so it goes
 
TheKing:
I actually just signed up for Code Academy (www.codeacademy.com) the other night. Not to learn R, but just to get my feet wet with some basic programming. Not really sure if I'll use any of it, but it seems like something we should ALL have a base level of knowledge in.

Anyone here ever use Code Academy?

No, but I would recommend udacity! They have a great cs101 class and the interaction tool is super cool.

 

Can anyone recommend a good way to brush up/learn stats?

It's been a while (and I was never that good to begin with), but I'd like to incorporate more statistical analysis into my day to day role in Corporate Finance. I tried a coursera class, but it was horrible. I don't know if any of the students actually finished the class.

twitter: @CorpFin_Guy
 

Both Udacity and Coursera have stats classes. The coursera one is passed I think, but the Udacity course can be picked up any time.

@Eddie - Thanks for this. I had signed up for a Couresera course on R (data anlysis using R), but missed a week and couldn't catch up. This seems like a trimmed version of the Coursera version, but will be a nice refresher.

 

@accountingbyday

I took Peter Norvig's Computer Science class on Udacity. I saw that Udacity also has an Introduction to Statistics course, which may be worth looking into. I don't know if the quality is comparable, but I've heard great things about CS101 and had a good experience with the class I took, so I assume it's probably pretty good.

At the very least, it's a free introduction.

(There's always reading, too! I wouldn't underestimate the power of spending some time with a solid book and working through some problems, assuming you have the self-discipline).

On an unrelated note, I'm a big, big fan of Udacity, because I think they're actually adopting education to better fit the Internet, whereas Coursera -- at least at the moment -- is simply putting educational content on the Internet. On the other hand, because Coursera courses are actually through Stanford, Berkeley, etc., they may have more sway when you put them on a resume. Admittedly, both companies are still in the very early stages of their development, so all these issues are eventually going to get hashed out.

 
atomic:
@accountingbyday

I took Peter Norvig's Computer Science class on Udacity.

Hahaha, Norvig's class! Y'know it was initially defined as a follow up to CS101 (which was literally a complete intro for someone who's never programmed before)!. I went through about 4 weeks of Norvig's class (which my boss recommended me to check out). I got absolutely wrecked every time by his solutions, but the stuff was so damn cool. Especially the world problem equations. I thought he was talking about solving one equation, like "ONE + ONE = TWO", and was shocked when he solved the entire class of problems!

 
canas15:
atomic:
@accountingbyday

I took Peter Norvig's Computer Science class on Udacity.

Hahaha, Norvig's class! Y'know it was initially defined as a follow up to CS101 (which was literally a complete intro for someone who's never programmed before)!. I went through about 4 weeks of Norvig's class (which my boss recommended me to check out). I got absolutely wrecked every time by his solutions, but the stuff was so damn cool. Especially the world problem equations. I thought he was talking about solving one equation, like "ONE + ONE = TWO", and was shocked when he solved the entire class of problems!

Oh my, I imagine the whole 'Take this after taking CS101' concept didn't pan out for some people. I had some prior programming experience, and I still found the class pretty challenging. Admittedly, it was my first time coding in Python (On a related note, if my first programming language had been Python instead of C++, I think I would probably be a developer today. It's that much more pleasant). I forget which Unit it was, but when they asked you to pseudo-implement regular expressions, I was like, 'Holy hell, this class is for real.' I remember... I would write this ugly, ugly code that -- with a lot of pain -- got to the result Norvig was looking for, but then he'd implement the same damn thing in two lines.

Still, definitely rewarding when all was said and done. Pretty amazing that less than a year ago these options weren't really available to anybody. There's so much educational content available to anyone who's interested these days, and, in my mind, this is the best thing that's emerged from the Tech Boom 2.0.

(Note to take anything away from those of you who are really passionate about checking in to places on Foursquare or something).

@Eddie

I saw the email! I'm very excited, and I'm hopeful that employers take to this. Imagine if people could reboot/advance their careers by showing some dedication in relevant MOOC's: I think it could really do a lot of good. I'll admit that I am slightly nervous. While I would absolutely love to hire people who'd demonstrated an independent interest in a particular field of study (and did well), I think the Powers That Be might prefer to hold on to the traditional means of recruitment longer than those in the MOOC Movement might anticipate or think is fair.

Still, I'm hopeful, and I'm glad they're choosing Software Engineering as their first career services goal, for a few reasons: 1. Software engineering is very much a 'Can you do it or not?' occupation. They're less focused on pedigree, and more focused on raw technical ability. (Obviously, though, software engineers have their own culture that you'll need to mesh with to succeed in the field). 2. They're very in demand at the moment, so employers might be more willing to take the risk on this new type of recruitment in a hot market environment. 3. This relates to 1 and 2, but it's much easier to verify that someone had actually benefited from the MOOC he took if it's in Software Engineering, as it actually relates to the job. More 'finance-y' courses still haven't seemed to bridge the gap between the real world and the classroom, at least from my limited experience.

 

@atomic Speaking of Coursera and resume inclusion and whatnot, I just got an email from them yesterday stating they've started a Career Services office. That's some next-level shit for a MOOC.

Agree about the content, though. It is almost identical to a classroom environment structure-wise, where Udacity seems to be a little more relaxed and work at your own pace.

 
IlliniProgrammer:
Eddie, it's a free version of Matlab for stats people.

That's about it.

so Matlab > R for the hopeful potential quant?

If your dreams don't scare you, then they are not big enough. "There are two types of people in this world: People who say they pee in the shower, and dirty fucking liars."-Louis C.K.
 
wolverine19x89:
IlliniProgrammer:
Eddie, it's a free version of Matlab for stats people.

That's about it.

so Matlab > R for the hopeful potential quant?

Both are pretty good to know. Some firms are cheap and won't get Matlab. Others get Matlab. Matlab does have a nice debugger. But for super quanty stuff, I think you would just use C++, to be honest.

 

I've been wanting to get into deeper stats analysis and this is a great launching point. Thank you for bringing it to my attention.

The error of confirmation: we confirm our knowledge and scorn our ignorance.
 

R is a great tool and has numerous applications with in the statistical and computational world. Although the potential problem with it becoming a "main stream" finance tool is the "plug and chug" functions. Many people in statistics and operations research use these same functions but they understand what is going on behind the scenes. If people do not understand the functions they are using it will end up being another black box application that half the desk does not and will not ever understand. In the same stroke one can also propose that clients will have a steep learning curve as well. P.S All of the functions are open source... meaning they can be inaccurate.

"Honey can you please pass the mashed pwntatoes"
 

Eddie - anything by Code Academy is worth looking at. I've spent time looking at their courses and have been using their Backbone.js lessons as a baseline before I start fucking around with node. Good call promoting anything from Code Academy. Then again, I really do think that any knowledge of programing needs to be tempered to what your doing in finance and you might as well do it the old fashioned brute force way with C/C++ as opposed to with anything fancy. Either way, I'll check out R. It might be interesting to play around with.

 

The learning curb for R is pretty steep. It took me a lot less time to learn SAS and SQL than R, which i don't i know as well as either sas nor sql still.

What you end up doing is this: "I need to do a t-test, how the fuck do i do that?" * spend 20-30 mins on the web finding the answer. Dick around with the syntax and do it wrong like 3-4 times for like an hour or so. Granted this happens for all programing languages but R syntax really isn't as intuitive to pick up as something like sql so this process happens much more often.

The main reason people like R is because it's free and opensource. It occupies within the analytics market the same area as linux for OS'. Linux always had an issue of bridging the gap between power users and power users who can code. I feel you will have this same issue with R. I had researched R GUI's and there are few decent ones but they are quite limited in the functions they can perform.

What ends up happening in most companies when buying stats programs is this "R is too hard and who's supporting it if it fails?" "SPSS sucks" "Matlab, maybe but its pretty expensive. Only approve it for the engineers" "Everyone uses SAS, it's pretty decent and they provide a shit load of support."

"SAS it is"

 

Slightly unrelated but still on the topic of MOOCs... does anyone know of a good course for Statistics? Preferably for someone who is looking for a refresher. Udacity has an Intro to Stats course but it's very slow and very broad; Coursera had one a while back that was not very quantitative. Was hoping for something more rigorous.

Currently: future neurologist, current psychotherapist Previously: investor relations (top consulting firm), M&A consulting (Big 4), M&A banking (MM)
 

Nobis quos architecto sunt amet. Nihil blanditiis molestiae id suscipit molestiae et voluptatem. Sed cum recusandae blanditiis rerum minima quia. Ipsum incidunt eius velit natus.

Dolorem quod et aliquid error non ipsa earum ipsam. Sed inventore quia laboriosam sunt quia accusamus.

 

Quisquam quidem beatae nam ea. A cum quibusdam ut nesciunt.

Sed omnis placeat ut in. Sapiente dolor est iste. Vero expedita doloremque vero quia. Et neque necessitatibus tenetur nulla numquam totam quam omnis.

Quia delectus sequi iure. Ab quos mollitia ut. Fugit aut temporibus enim.

Currently: future neurologist, current psychotherapist Previously: investor relations (top consulting firm), M&A consulting (Big 4), M&A banking (MM)
 

Et nam dolore ea sit quia. Maxime dicta a sit sed error non accusantium. Ea et non et porro dignissimos. Animi consequatur rerum unde molestias. Sint iste et nisi voluptatem aut sit et. Est id quia harum consectetur sapiente voluptate consequatur.

Laborum vel vitae nostrum vel et. Aut ab optio atque commodi aut rem totam in. Autem quam et distinctio consequatur repellendus ducimus accusamus. Id possimus quam ad dolores sunt inventore. Nihil fuga quasi aliquam magni ad et quod. Tempora omnis vitae velit magnam sit aut voluptas. Veritatis voluptas debitis accusantium impedit.

A laboriosam et accusantium et velit magnam. Iste aliquam rem repudiandae voluptatem minima sequi.

Career Advancement Opportunities

March 2024 Investment Banking

  • Jefferies & Company 02 99.4%
  • Goldman Sachs 19 98.8%
  • Harris Williams & Co. (++) 98.3%
  • Lazard Freres 02 97.7%
  • JPMorgan Chase 03 97.1%

Overall Employee Satisfaction

March 2024 Investment Banking

  • Harris Williams & Co. 18 99.4%
  • JPMorgan Chase 10 98.8%
  • Lazard Freres 05 98.3%
  • Morgan Stanley 07 97.7%
  • William Blair 03 97.1%

Professional Growth Opportunities

March 2024 Investment Banking

  • Lazard Freres 01 99.4%
  • Jefferies & Company 02 98.8%
  • Goldman Sachs 17 98.3%
  • Moelis & Company 07 97.7%
  • JPMorgan Chase 05 97.1%

Total Avg Compensation

March 2024 Investment Banking

  • Director/MD (5) $648
  • Vice President (19) $385
  • Associates (86) $261
  • 3rd+ Year Analyst (13) $181
  • Intern/Summer Associate (33) $170
  • 2nd Year Analyst (66) $168
  • 1st Year Analyst (202) $159
  • Intern/Summer Analyst (144) $101
notes
16 IB Interviews Notes

“... there’s no excuse to not take advantage of the resources out there available to you. Best value for your $ are the...”

Leaderboard

1
redever's picture
redever
99.2
2
Secyh62's picture
Secyh62
99.0
3
Betsy Massar's picture
Betsy Massar
99.0
4
BankonBanking's picture
BankonBanking
99.0
5
dosk17's picture
dosk17
98.9
6
DrApeman's picture
DrApeman
98.9
7
kanon's picture
kanon
98.9
8
CompBanker's picture
CompBanker
98.9
9
GameTheory's picture
GameTheory
98.9
10
Jamoldo's picture
Jamoldo
98.8
success
From 10 rejections to 1 dream investment banking internship

“... I believe it was the single biggest reason why I ended up with an offer...”