General feedback/various points

Module 1. The Predictive Modeling Pipeline Tabular data exploration
#1

Ahoi hoi folks,

as I have various smaller points I would like to mention/outline, I thought creating a comprehensive post would be better than creating one for each. However, if the latter is preferred, I can of course also adapt things accordingly. Please just let me know.

  • cell 4 maybe mention that dataset is restricted to certain features
  • cell 10 maybe add information wrt redundancy to note
  • cell 11 maybe add note that plot is restricted to 5k observation to speed up plot and keep it readable (I assume these are potential reasons)
  • An example of machine learning model decision rules
    • graphic: x label cut off, maybe a bit confusing as class labels and probability have the same color (range)
    • paragraph below bullet points: “…in a high-dimensional space …”, dot missing at end of sentence

I hope the points/comments are understandable. If not, please let me know if you have questions.

HTH, cheers, Peer

#2

Thanks a lot for your feed-back! Could you try to add snapshots (Discourse makes this quite straightforward using copy and paste)? Cell number is not super useful for various reasons (for example I need to reexecute the notebook first, and we may not have exactly the same notebook version).

cell 11 maybe add note that plot is restricted to 5k observation to speed up plot and keep it readable (I assume these are potential reasons)

Those are the two reasons indeed IIRC.

An example of machine learning model decision rules

This section we will probably tweak in order to simplify it see First look at our dataset for more details, which means that your comments will automatically be “fixed” (or at least not apply any more) when we change this section.

#3

Ahoi hoi @lesteve,

thx for the quick reply. Wrt cells: ah sorry, I forgot to add that this refers to chronological order of code cells and not e.g. run states and I didn’t know that there are different versions, sorry! I’ll try to add this to my other posts.

Cheers, Peer

#4

The example of machine learning models will be changed and I fixed the remaining points.

Don’t worry about the cell numbers it is fine after all, I just need to run all the cells in a notebook and we did not change that much the code cells so the numbers are accurate in a vast marjority of cases.

#5