Wrangling chapter review

[leem44]

@trevorcampbell @ttimbers Please have a look at this PR and let me know if I should make any changes!
This PR addresses: #103, #160

• Note: I have not yet removed map from this chapter -- I figure its easier to get any feedback on this, make changes and then create a branch off of this to remove map

[leem44]

@ttimbers Changes look good! Only thing is the one below "Montreal" is spelled "Montreak"

Also changed the vector figure to be of type character, as previously it was called an integer, but then we actually made a double vector in code... I tried to swap it to double, but then it was too confusing for a first example. Character vector is a simpler and thus better first example:

[ttimbers]

Good catch!

[ttimbers]

I also removed loading data from the canlang package and instead get them to read the data from files. I don't think we really explain data packages and they are kind of weird, so best to leave out I think?

[leem44]

Yes, I agree with that! Makes sense!

[ttimbers]

In Fig 3.10 we illlustrate going from longer to wider, but each table is the same number of columns. Yes, this is technically something that happens, but for our first teaching case we really should have a wider dataframe that we make longer:

Also, where did the values for commuters come from? I imagine population came from canlang, but I don't think that data set has commuters? Other columns we could add from canlang could be area dwellings or households (we need something that is a count to stick with the narrative)?

[ttimbers]

I think we should remove the & operator from here as we never use it elsewhere in the book or course. So it's just redundant and extra info that they do not "need" to know:

[ttimbers]

@leem44 - I am also removing the demonstration of filter not working on numeric data that is in the mutate section, as it is rendundant to what we wrote above in the section about convert and separate.

[ttimbers]

@leem44 - in the wrangling chapter, are you OK to move the pipe to before mutate and after select and filter? It uses select and filter so is timely after those two. Also, then I can use the |> in the setup for the as.numeric transformation for mutate so the creation of the data frame with the character columns that should be numbers is a little less magical?

If I do that, then I will use arrange in the pipe preamble since that was intro'ed in chapter 1, but mutate won't have been demo'd yet. Also, I think that works nicely since we use arrange in the code example we show that follows.

[ttimbers]

Maybe not actually - I am now fence sitting on this change, as it causes cascading changes...

Decided not to do this.

[ttimbers]

Simplified the example to just plot the proportion of people speaking English as the primary language at home.
Then we can skip case_when and vector recycling and keep the narrative pretty similar.

[ttimbers]

Done all sections except for summarize + across and purrr::map* (I did do rowwise which comes after these).

[ttimbers]

OK, I think we are ready to merge. I added a bunch of images into the wrangling chapter, and re-organized the section where we do aggregation. I am pretty excited about it. @leem44 - I made quite a few changes to the wrangling chapter to (hopefully) simplify it. I hope that is OK. If there is anything I removed that you really want to keep in, please let me know and we can talk and find a way to keep it.

[ttimbers]

@leem44 gave me the go-ahead to merge, so merging!