Feedback Module 4

[hfboyce]

Hi @mgelbart,

The first draft of Module 4 is available here. I am less happy with the sections on operators (18 and 22) so feel free to give any ideas you may have to make them better.

The whole module has 26 exercises.
We made some changes to the docker container so I am hopeful that you may have less waiting time with them now (< or ~ 20secs).

[mgelbart]

• 1.3: There are -> Here are some data types built-in to the Python language
• 1.4 quotation -> quotation marks; extra *; positive, negative, or zero; classified -> classifies
• 1.7 capital "A"; "but None is an actual data type"; maybe add a summary sentence: "In summary, NoneType is its own type, with only one possible value, None; on the other hand NaN is actually a special float value. Pandas dataframes tend to use NaN for missing values."
• 1.8: stating -> indicating
• 1.9: called a string (str); enclosed -> can enclose; use code font for the examples like 'Hello';
• 1.10: if you do print(saying) it will format properly, is it worth showing that?;; "When we discussed None type and we didn’t yet know the monster’s name, why didn’t we just put empty quotations?" -> hm, well that does seem like it could have been reasonable. Maybe just say an empty string is sort of like NaN in that it has a type but no data?
• 1.13 "Sometimes we need to explicitly change the type of an object; this is called casting." In general, let's make it more clear you can cast objects, not just values.
• 3: if you use triple quotes you won't need the \ at the end of every line - would that be less confusing?; learn->learned; in the hint, tell them to convert the entire lyrics to lower, that might not be obvious; Binder/Docker not faster for me, unfortunately
• 4.1: you actually can, so long as the string is something like '4' or '0'. So the question/answer is misleading.
• 5.2: Capital "L"
• 5.3: Capital "L"; "The elements in a list can be a combination of any of the data types." -> should we show that the elements can also be lists, or is that overboard? Update: ✔️
• 5.4: Let's start with the length of the simpler (non-nested) list, then do this one. And for this one, let's say it's the number of elements in the outer list, not the total number of elements.
• 5.5: I'd do all this with a non-nested list to avoid confusion.
• 5.7: too complicated?
• 5.8: remove comma in last sentence
• 5.9: "list have" -> "lists have"; "its data" -> "their data";
• 5.10: show the type of shopping_items
• 5.11: "They can be contained in parenthesis" -> "They are represented with parentheses"; show casting a tuple to a list?
• 5.12: "Sets, not unlike lists and tuples is a data structure that contains elements" -> "Sets, not unlike lists and tuples, are a data structure that contains elements"; formatting error with bullet points;
• Pls go through for capitalization and punctuation at the end of sentences; I won't list them all.
• 5.13: "Remember this data structure has no order, so Python displays them according to some internal sorting scheme".
• 5.14: sets are mutable
• I think we should show concatenation with + for str, list, tuple somewhere
• 6.1: missing the correct answer?
• 7: Bug in the code, the buttons for 7.2 seem to be linked to 7.1
• 8: something is broken

[mgelbart]

• 10.2: formatting issues with **; could we go with some shorter keys/values to make it clearer and the colon stand out more? the house example from the next slide is better.
• 10.3: it's only the values that can be of any type; "are unique" -> "cannot contain duplicates"
• 10.4: have we already talked about mutable vs. immutable at this point? You can say that immutable types can usually be keys and mutable types usually can't.
• 10.6: do they need .keys() and .values()? maybe we can delete this slide, drop keys/values, and teach .items() when we teach loops?
• 10.9: typo "we we"
• 10.10: delete this slide?
• 10.11: remove the "add values" column? Also, maybe we want to change "Ordered" to "preserves order". Sets do have an order to them, I guess, but it's just not necessarily the same order as you put them in. E.g. {1,3,2} turns into {1, 2, 3}. We might want to also have a slide explaining this idea of preserving order; let's also change "can accept duplicates" to "can contain duplicates" - a set can technically accept them as in you can type {1,2,2} but it will not keep them both.
• 12: feels a bit repetitive, maybe reduce the number of entries?
• 13: same comment as 12
• 14.3: * formatting issues? also no need for newlines after the colons
• 14.6 link formatting error; capital "L"; typo "the axis labelled"; also, is axis the right word here? maybe the series?
• 14.7: lower case "C"; remove "but there are others"
• 14.11: delete?
• 14.12: "These" is confusing - change to "Non-numeric types"
• BTW, will be talk about how pandas automatically tries to recognize types in read_csv, but it sometimes fails? And I think you can provide type info to read_csv maybe? Also, df.info() might be relevant here? or df.describe() I forget which, one of them shows types. or maybe df.dtypes is good enough. UPDATE: some if this shows up in Ex 22.
• 15.2: extra "Who am i?"; also, here I'd give the full name pandas.core. whatever, instead of the abbreviation
• 16: Same bug as earlier on, the two quetsions are linked somehow. Also 16.1 has both options as False.
• 17: this is a bit boring. Any ideas?
• 18.2: I think i mentioned this earlier, but I suggest we remove // from the course. can we also get away with removing % or do we need it somewhere?
• 18.3: remove //
• 18.4: this slide is missing the underlying process: that True gets casted to 1 and False gets casted to 0.
• 18.6: though, you can multiply strings by integers - seems worth mentioning here; you said "transform", did we introduce the term "cast" anywhere? did we teach casting via str() yet?
• 18.7: supportive -> supported
• 18.8-9: it's feeling a bit dry and repetitive by now. Can these be moved to exercises where they dry stuff out and answer MCQs about what happens?
• 18.11: comma after and; the != makes it quite confusing, can you change this to 6<7 and dogs == cats ?
• 19: don't need code font for "error". also, change "Output" to "Runs successfully" or something like that; delete 19.3 if we're deleting //
• 20.1: I don't think we need any of the stuff above the text.
• 20.2 and 20.3: though not strictly needed, I'd feel better with parentheses, like not ('hotels' != 'homes')
• 21: feels like we're getting a bit far from data science here - maybe delete this exercise? we had some good stuff in the past relating to indexing dataframes with and/or, could we bring some of that back instead?
• 22.4: change period to comma
• 22.4: make clear that this is because it's a string, and we already learned that summing a string results in concatenation!
• 22.8: I found this surprising, I would have expected max to return the last one in alphabetical order. Are you sure it's the most frequent?
• 22.9: this is a useful but non-trivial point. some more scaffolding would be good. maybe exercises and more explanation.
• 22.10: clarify wording to: "axis=1 refers to the calculation being done for a row, across multiple columns, whereas axis=0 (which is the default for aggregation verbs) refers to the calculation for a column, across multiple rows."
• 22 is starting to feel pretty heavy overall, I wonder if it needs splitting.
• 22.12: is displaying both the manufacturer (N, Q, etc.) and the cereal type (Cold, Hot).
• 22.13: what happens if expand=False ?
• Nice job first showing the output in 13 and then putting it back into the df in 14-15.
• Though, there's probalby a way to do it without creating an intermediate df like new, right? this feels a bit clunky. Meh, it's also fine as it is.
• 23: really like these
• 24: are these new skills from Module 4? They feel like old news, but maybe I'm confused.

[mgelbart]

(first wrote this in VS Code, then copy/pasted here, to be safe 😬 )

[hfboyce]

Just a couple of small questions:

10.4: have we already talked about mutable vs. immutable at this point? You can say that immutable types can usually be keys and mutable types usually can't.

I have talked to them about this but Im not clear of what you mean here.

17: this is a bit boring. Any ideas?

IT IS!!! I've made it a little different. It's still a little boring, but maybe less so?

21: feels like we're getting a bit far from data science here - maybe delete this exercise? we had some good stuff in the past relating to indexing dataframes with and/or, could we bring some of that back instead?

Which stuff? What are you thinking here?