The first MOOC I completed ended on February 10th. Foundations of Data Analysis, a course by the University of Texas at Austin was hosted on edx and essentially, was an overview of statistics from regression to hypothesis testing. The course itself was excellent at introducing concepts and putting forth the theory while at the same time it introduced R, a statistical analysis tool. Personally, the material was a review for me as I had taken a statistics course, as well as an econometrics course during my undergraduate studies. That was a few (11 since graduation) years ago and apart from making you think differently there has not been any practical application of this knowledge.
Each lesson was released on a Tuesday consisting of a reading (pdf), lecture videos, R tutorial videos, pre-lab, lab and problem set. The readings were chapters from a statistics textbook and provided the foundation of the theory for each topic. Lecture videos was additional theory and insight on the topic by Dr Mahometra, while the R tutorial videos introduced the related R commands. The pre-lab and lab were a series of exercises were you could apply the theory and R commands. The lab and pre-lab were very similar in nature; basically the pre-lab was guided by the videos while the lab was work you had to complete unguided. Finally, there was a problem set at the end for additional practice.
Statistics adds a lot to the way you view data and information, either at work or even in the news. However, unless you are directly working at a company that analyses data in this way you will not come across most of these concepts. You might use some basic descriptive stats but inferential statistics as well as modelling are rare. In RA specifically, statistics will be used above and beyond day to day activities. They could be used initially to “sniff out” some characteristics of the data, but the inferential and modelling parts will come only at a stage when everything is running so smoothly that RA is providing additional information from the checks demanded.
For anyone wanting to get a good overview of statistics this course is ideal. The fact that you also learn a modern statistical tool in the process is an added benefit. My only reservation is not specifically with this course but with MOOCs and training outside of working in general. Depending on your stage in life your time can be scarce and this means that you need to be selective with how you choose to spend it. There will be times when you might feel like the knowledge you gain is not immediately applicable and that it is not worth it. However, another way to look at it is that you are adding to and building your resume as well. You are building for the future or you can simply be reviewing. If you deal with data on a daily basis – as most RA professionals do – then some knowledge of statistics would help you even if it’s not something you can use right away.
The next courses I have registered to take are:
1. Communicating Strategically – PurdueX – started this week – duration 5 weeks – “Improving communication skills and presentation effectiveness for scientists, engineers, and professionals.”
2. Networks, Crowds and Markets – CornellX – 16 Feb 2015 – duration 10 weeks – “an exploration of fundamental questions about how our social, economic, and technological worlds are connected. Students will explore game theory, the structure of the Internet, social contagion, the spread of social power and popularity, and information cascades.”
3. The analytics edge – MITx – 3 Mar 2015 – duration 12 weeks – “apply analytics to real-world applications” – looks like the most challenging but also the most relevant and rewarding one for an analyst.
4. Introduction to Big Data with Apache Spark – UC BerkeleyX – April 2015 – duration 4 weeks (python or other programming language prerequisite) – “Learn how to apply data science techniques using parallel programming in Apache Spark to explore big (and small) data.”
For each course, I will usually decide if I will follow it to the end during the first week. This will also depend on my time availability of course… For the Data Analysis MOOC I added pressure to myself because I wanted to see it through so I paid for an ID verified certificate. I crammed at the end to achieve the required score to get the certificate, but it was worth it. If nothing more than for the sense of accomplishment.