Lab 08 Issues

[charlesfrye]

Too long! Less coding next time.

[charlesfrye]

~"add the mean and multiply by the standard deviation" is wrong order of operations in description of how to generate noisy data~

[charlesfrye]

~Does .from_dict appear earlier?~

~Folks had trouble with the section where they needed to make a pandas dictionary. This should be more clear.~

[charlesfrye]

~formula is wrong for correlation!~

[charlesfrye]

~inputRange for visualization of CDF as function of sample size should be -5,5, not 0,5~

[charlesfrye]

~07b is missing the header image~

[charlesfrye]

make sure to include a plt.figure in each cell where you're making a plot.

->

make sure to include a plt.figure() in each cell where you're making a plot.

E 10/11/2017: Done

[charlesfrye]

In the section where randint is explained, students aren't actually asked to do anything.

[charlesfrye]

~Students should be directed to use no more than 30 samples when making their own plot of the CDF , so they can more clearly see the connection to the rugplot.~

[charlesfrye]

Clarify the meaning of "average" for "average CDF" by comparing it to calculating the average of numbers.

[charlesfrye]

For resampling: jump from random.choice to random.randint is confusing. Though it's less efficient, random.randint works for both and so should be used for both.

[charlesfrye]

~Show an example of using plt.figure() so that folks know what to do.~

[charlesfrye]

~Alternatively to having students plot the eCDF themselves, could introduce plt.step~

[charlesfrye]

~Remove challenge problem with lambda? No one appears to have attempted it.~

[charlesfrye]

These labs had too much material and were too challenging. They also need to be restructured in light of the inclusion of weekly tutorials. The following changes should fix.

• [x] Include two tutorials: one on bootstrapping and one on correlation.
• [x] Move sections on randomness and resampling to bootstrapping tutorial.
• [x] Add an animation that directly compares the eCDF for bootstrap samples with the eCDF for resampling from the original distribution.
• [x] Add section on bootstrap hypothesis testing for one-sample tests to Lab A.
• [x] Move the section on nonlinearity and correlation to the tutorial on correlation.

[charlesfrye]

finished!