I recently decided to learn programming in the R language, when looking for a class to help me learn I wanted to choose something that would not only teach me the basics of the language but also how to us it in real world situations. I chose to take a look through some of the Massively Open Online Classes (MOOCs) that are available. Below is a quick review of 3 classes that I took to learn R.

Analytics Edge by MIT edX

This class was by far my favorite. It really inspired me to want to learn more and pursue things outside of the class. The class is being offered again this March and can be audited at any time while it is in an archived status, so you should be able to take it any time you would like.

This class does not require a lot of prerequisite knowledge, just some basic statistics. It does, however, require a fairly large time commitment, requiring a minimum 10 hour per week to watch the lectures and complete the assignments.

The instructor would use interesting, real world examples to present the concepts, then guide you through the process in R. At times the R tutorials were a little too "type this - get that" as opposed to actually having to think about what you are typing, but the lectures were easy to follow and very helpful. One of the best parts of this class is that one of the assignments is to enter an actual Kaggle competition. For anyone unfamiliar with Kaggle, it is a site that hosts data science competitions, some for fun and some for real cash prizes.

I also really like the edX platform, it's easy to know where you are and where you are going within the class.

Coursera Data Science Specialization

The Data Science specialization from Coursera is composed of 9 Courses and a capstone project that teach a broader range of skills than the other classes do. While it is 9 different courses it really feels like one really big class with 9 parts. Each class is one month long, but you can take 2 or even 3 at a time. I mostly took 2 at a time to get through it faster, I tried 3 once but that was a little too time consuming if you are working full time.

There is a paid "Certified Track" but I found that the free versions of the courses gave you the same level of instruction without any of the cost.

My complaint about this series is that sometimes the lectures were difficult to follow, I found myself looking for other videos to explain the same topics, and often found much better explanations from other sources. The classes make a pretty ambitious goal of trying to teach very large topics with just a single short MOOC. Specifically the Statistical Inference class attempts to run through many complex statistics concepts very quickly and I don't think it does any of them justice. If you are going to take this series, learn basic stats somewhere else.

Also, while they seem to quickly cover some of the more core concepts they spend quite a while teaching things that I personally can't see as being as valuable. While using R to create a presentation (rpres) is pretty cool it doesn't seem like someone who is participating in this series would really need.

One big plus for this series is the sheer number of people taking it. The community was very helpful through the forums and even a group I found on LinkedIn. Also, when you are finished with the series you have quite a few files that you can use to help create a professional portfolio to show of your newly acquired skills.

Statistical Learning from Stanford edX

If you have a bit of Stats and Math background, you may want to consider Statistical Learning from Stanford edX. The professors say that this is not a Math heavy class, but they certainly go more in depth there than the other courses I took. It does include an introduction to R and provides a good starting point for folks who have never used it, but the benefit of this class is also the instructors. Here you are getting a couple of professors that are some of the thought leaders in statistics and they even include interviews with folks like John Chambers (inventor of the S language which is the predecessor to R).

Sometimes the lectures aren't as inspired but this class may give you the best information and the best understanding of the topics once completed. The textbook is also a great resource, you can download an electronic copy for free, but I opted to get the real thing. I've actually referred back to that book several times since the end of the class.

The workload is not too tough, probably about 5 hours per week, but again this assumes that you already have the basics down.

I will also mention a couple of resources if you are just getting started with R. Try R and Datacamp's Introduction to R are both pretty good for those just starting out. These are your typical introductions to a new language, they don't give you a lot of real world examples but do teach some of the basic syntax and conventions of the language. I would advise using one of these along with one of the other courses.