Thursday, January 10, 2013

Coursera Computing for Data Analysis, Week 2

Coursera Confusion surrounds the week 1 programming assignment (which was just a set of MCQs). At least the instructor is trying his best on the communication front:
"Unfortunately, my attempt to fix Question 9 last week caused more problems and confusion than it fixed. Therefore, I've decided to delete the original Question 9 from the Quiz. Now the Quiz for the first Programming Assignment only has 9 questions total instead of 10. I have adjusted the grading for this Quiz and have issued a regrade. If you have already taken this Quiz, you should NOT have to take it again.")

It's clear the Coursera does a crap job of road testing units before they offer them to students. Anyhow, on to week two.
"A few facts about the class so far. There are currently 40,211 people enrolled in the class and about 31,000 active users in the past week. There were about 14,000 who submitted the Week 1 Quiz and about 1,300 people participating in the forums."

The course was a bit rough this week, (Control structures, Functions, Loop functions, Debugging) - full on programming, no statistics. I'm going to stick it out for for one more week because I want to get to week 3 (Simulation, Plotting, Visualizing data, Principles of data graphics). After that I will bail because this module is about programming not stats so it doesn't really fulfill my learning objectives, and because I'm away the following week. Also, Coursera Data Analysis and Google Advanced Power Search both start that week and they will take up all my time.


  1. I'm taking this course also...

    It's my third Coursera course. The first was very well organised, the second total rubbish (though I stuck with it and still learned a lot). This seems to be somewhere in between so far though I'm hopeful it will get better.

    I get the sense that quality control is up to the individual offering the course. (I assume that their affiliated institutions are largely unaware of what is going on and therefore don't provide much pressure either). Presumably this is something that Cousera will want to change.

    I've primarily taken the courses to learn different programming languages so far - I'm not that interested in the topics. The first was about SaaS (I learned Ruby). The second about financial investing (I learned Python). This is about stats (notionally) - I'm learning R.

    In short, I see the courses as little more than an excuse to make me do something.

    So far, I'm happy with that approach.

  2. I agree. That's why I have found it necessary to set my own learning outcomes for MOOCs, which don't necessarily align with the "official" set. MOOCs as smorgasbord - let us graze.