AaltoSciComp / python-for-scicomp

Python for scientific computing course
https://aaltoscicomp.github.io/python-for-scicomp/
Other
25 stars 31 forks source link

Ideas for plotting exercises #37

Closed bast closed 4 years ago

bast commented 4 years ago

Currently:

Suggestions:

Add exercise:

bast commented 4 years ago

For the CSV examples (learner chooses one):

bast commented 4 years ago

Idea for 4.1: provide data for https://seaborn.pydata.org/examples/anscombes_quartet.html and let learners add a linear regression. This data set is a classic and used as an example why looking at data alone without plotting it can be dangerous.

samumantha commented 4 years ago

Agreed, the datasets used for plotting should be as simple as possible. Stateful porting to oo is a good exercise since when googling problems both come up mixed and things do not work if you don't know about the two ways of doing things. As something geospatial, maybe plotting a tif, like digital elevation model ?

bast commented 4 years ago

I admit that for the longest time I did not even know there were two interfaces to matplotlib (I know I should have just opened the page and read the doc for 5 minutes) and I was many times confused why examples always looked differently than I remembered.

samumantha commented 4 years ago

same here.

annefou commented 4 years ago

I like your plan @bast You probably have it but is there some details on how to customize your plots (changing font size, labels, etc.). I see too many researchers importing png file in powerpoints and overwritting labels, titles. Not so good for repeatability/reproducibility.

bast commented 4 years ago

Good point, I will add such a section.