I've taken this course, and while the instructors clearly put a lot of effort into it, I found it missed the mark by quite a lot.
The assignments use fancy algorithms, but often with little motivation; as such, I doubt I will remember most techniques. Furthermore, the class didn't have enough prerequisites, and so the derivations and discussions were often soft or missing --- again, making it harder to learn. Most importantly, the time you spend on the class is dominated by the homework, but it suffered from the issue a lot of online coding courses have: if you are told to write a function, given its signature, behavior, etc, you aren't learning the material. instead, you chug away at it like it's programming 101, pass the tests, and move on without reading up on the methods that used that function. and that's exactly what happens with a large portion of the homework assignments in this class.
+1
this course was very much so either
a) Do this basically exactly how we told you do do it, basically "filling in the blanks"
b) Predict the midterm elections (well fuck, this is an intro class, I'm really not sure what to do)
I found it to be quite poorly structured, and its lessons did not translate well into being able to do data science off on your own problems
Sounds symptomatic of many programming classes catering to non-programmers. The better programmers don't have enough material to go in depth, while the weaker ones get enough paint-by-numbers to get through, without really understanding anything.
And for better or worse, so much of data science is cleaning data and figuring out how to link datasets that don't have good indexes.
Does anyone know the name for graphs like "Station Fullness Over Time of Day" is?
I don't remember having seen one before (I don't look at many graphs) but find it incredibly readable, even being color deficient. Also, can't say I dug further than the homepage, but it was striking enough for me to ask.
I took this class last semester and our team won best overall project from approx. 100 teams. iPython notebooks and writeup can be found at http://project.chanceme.info/
I live in third world country, and universities are ludicrous. I have said this repeatedly, but I want to say that again because education is the important thing in the world in my opinion, and free lectures/education is just like discovery of fire in human history.
If I were to choose prophets for humanity, I would choose the people who provide free scientific material for humanity.
Not a bunch of schizophrenic from 14 or 20 century ago.
For those who live in developed country maybe you cannot realize how much impact can a simple series of lectures can have on students on third world country. I live in third world country, I can see with my own eyes how can it change people mind.
The assignments use fancy algorithms, but often with little motivation; as such, I doubt I will remember most techniques. Furthermore, the class didn't have enough prerequisites, and so the derivations and discussions were often soft or missing --- again, making it harder to learn. Most importantly, the time you spend on the class is dominated by the homework, but it suffered from the issue a lot of online coding courses have: if you are told to write a function, given its signature, behavior, etc, you aren't learning the material. instead, you chug away at it like it's programming 101, pass the tests, and move on without reading up on the methods that used that function. and that's exactly what happens with a large portion of the homework assignments in this class.