Class is in session. Columbia University's Machine Learning for Data Science started last week. After just one week, I have to say this class is very different from the many other MOOCs that I have taken so far.
For starters, Columbia is offering three separate data science classes: Statistical Thinking for Data Science and Analytics, Machine Learning for Data Science and Analytics and Enabling Technologies for Data Science and Analytics: The Internet of Things. If you complete all three courses, you qualify for the "XSeries" certificate. Who doesn't want a certificate that starts with an "X" in it? I know Wolverine would.
Anyway, its surprising that for the Machine Learning class there doesn't seem to have any programming component, which is rather surprising for a data science course. I have already taken several data science classes from Johns Hopkins and they all required some programming. The Machine Learning course just requires students to answer multiple questions directly related to the video.
The other unusual thing is that the material moves quite fast. In week 1, we already covered merge sort and Big-O notation. MIT's Introduction to Computer Science and Programming Using Python didn't cover Big-O until a couple of weeks in. In the private course discussion forums, there is a lot of confusion regarding seemingly simple things. Its like Columbia designed a course to make a technical subject accessible to non-programmers although it really requires someone with a technical background. Sound like a typical MOOC problem.