reviewer comments (pdf)

[x] p.13 move info box from page 20 to start of chapter will make explanation clear
[x] p.16 give title of book "Flatland"
[x] Fig 1.5 clarify: Is the data from uniformly random points in a ball in Rd projected onto 2D orthogonal axes?
[x] p.23 I would be careful using Xj to specify a row, since this could equally be used to specify a column. Instead, you could use Xj. and X.j to make the column and row distinction clear.
[ ] p.27 The description in the box of “low-dimensional” space is not very good, since it seems to suggest that a lower dimensional space is just an interval. Likewise, the sentence in italics on this page seems to be a bit misleading. If you project the data onto 2-dimensions then it is still 2D, even if the variance in one of the dimensions is reduced.
[ ] p. 37 The description of PCA could be a bit clearer. I would say that you first find the projection that captures the most variance, then you find a projection that is orthogonal to the first axis that captures the most variance, etc.
[ ] p. 57 The notation in the equation on this page could be better to make more consistent use of subscripts; e.g., I would use $d_p(i, j)$ and $d_k(i, j)$ or something like that to show that they are both measuring distances, but just in spaces with different dimensions.

CODE

[ ] make sure code_and_data.zip file link is correct
[ ] p.28 It might be good to include the code that actually creates the plots for figure 3.1. Otherwise, I'm not sure it's that useful to have the code that creates the data.
[ ] p. 29 There is a deprecation warning for the code at the bottom so it would be good to update the code to work with ggplot 3.4.0
[ ] p. 30 There is also a warning when running the code at the bottom of page 'All aesthetics have length 1,'... I recommend running all code in the book with the latest version of R to catch any warnings.

Ch 4

[ ] p. 41 Including the following line to create a directory for the gifs is needed before running the code: dir.create("gifs"). (Give instructions to create the gifs folder.)
[ ] p. 42 The code from the call to the render_gif() function did not render as an animated gif on my mac when viewed using the default "preview" application but instead rendered as a sequence of still images. DIDN'T HAVE GIFSKI INSTALL, PBLY - ALSO MAYBE NEED TO NOTE ABOUT RSTUDIO RUN IN CONSOLE TO GET DISPLAY IN GRAPHICS WINDOW
[ ] p. 43 the code did not render at all and stalled with the message saying "Converting input data to required matrix format" (I let the code run for 20 minutes before pressing Esc to quit)
[ ] p. 59, 61 The data frame cnl_umap is not defined so I can't run the code on these pages.

dicook / mulgar_book