Interesting but often difficult to follow

M4. Linear Models M4. Wrap-up quiz 4
#1

Dear all, in general the quizzes are very useful, but
sometimes the burden of apparently simple tasks (which are not the focus of the mooc), like plotting or simply getting names of columns in the right format is making difficult to full profit from the quiz itself as I spend a lot of time to understand why for example in creating a Panda Frame columns=data.columns is not good or why the shape is not the correct one.
If possible I would add more tips on not-main tasks and highlight the tips very early in the quiz text.

Hoping this can help and thanks again for this very interesting and useful mooc!

Best regards,
Marco Bomben

1 Like
#2

I do get your point, if you have some specific quiz questions where a hint would have helped a lot, please do tell us! This way we can try to improve for the next session!

#3

Dear @lesteve , all,
for example at question 10 I had to look at the solution as I could not find a way to explore the results of the cross-validation step.
What was blocking me was this line:

coefs = [pipeline[-1].coef_[0] for pipeline in cv_results["estimator"]]

I still don’t understand why .coef_ in this case is not “enough” while in previous questions it was it and coef_[0] is necessary.
As I was saying in my original post I understand it is a minor thing with respect to the goal of the mooc but this issue blocked me in studying the results.

Many thanks again for your help!

Best regards,
Marco Bomben

#4

I would agree. The peculiarities of Pandas make going through these quizzes a lot more difficult than the actual content of the module.

Maybe the addition of commented code snippets as optional hints for how to wrangle the data might be a nice addition for future sessions (e.g., how do we get the coefficients again or how do we concatenate the resulting arrays efficiently for plotting).

1 Like