Meeting chat log
00:07:43 Daniel Chen: hi Jon. hi everyone. first time here :)
00:08:03 Jon Harmon (jonthegeek): Welcome!
00:09:27 KakShA Ekam DallasSK: the gang is here..except for tyler
00:09:31 KakShA Ekam DallasSK: and his kid
00:10:01 David Severski: The 4 seasons of hex bin stickers. LOL
00:10:11 Tan Ho: MiniTyler doesn't need help with the tidyverse
00:10:44 KakShA Ekam DallasSK: that's right..he just wanted to learn bangbang
00:11:11 Tan Ho: hey, there's Tyler!
00:11:23 pavitra: yay!!
00:12:36 Daniel Chen: i feel like more people have cameras on and respond here than any other zoom meeting i've been in all year...
00:12:50 Tony ElHabr: haha I second that
00:13:03 Tan Ho: we're all frens here :)
00:17:25 Tony ElHabr: i will not hate on using mtcars i will not hate on using mtcars i will not hate on using mtcars
00:17:30 Tan Ho: Ooh! A use case for the <details> block
00:17:41 Tan Ho: shush, no dataset snobbery
00:17:54 Darya Vanichkina: Yup, details is so helpful for easily hiding stuff in Rmd
00:18:06 Jon Harmon (jonthegeek): Ooooh, good call on details, let's make it clear how to do that in the instructions! (I'm still partly here).
00:18:25 Yoni Sidi: Shameless plug {details} https://github.com/yonicd/details
00:18:48 Tan Ho: ohhh dang
00:18:53 Tan Ho: installing now, I use that all the time
00:19:24 mayagans: Ooooooh it makes a collapsible div?? That is sick nasty!
00:19:29 pavitra: wow, details is very cool
00:19:34 David Severski: Funny, I was just exposed to details in GH issues earlier today. :)
00:20:03 Yoni Sidi: Seamless integration to rmd with it’s own knitter chunk engine…ok I’m done plugging
00:20:15 Vasant M: Ah yes details! That’s why I recall your name @yoni Sidi
00:20:31 Tony ElHabr: lol i have also been having this weird data frame print out issue
00:20:55 Darya Vanichkina: I wonder why…
00:21:27 Tan Ho: must be new tibble package related issue
00:22:17 Yoni Sidi: What print problem?
00:22:22 Daniel Chen: fyi: not all things that work on tibbles will work on data.frame
00:22:36 Tony ElHabr: things like "[90m" get printed out instead of "#"
00:22:53 Gabriela Palomo: and viceversa data.frame -> tibbles
00:22:55 Tan Ho: https://r4ds.github.io/bookclub-tmwr/reusing-existing-data-structures.html
00:23:39 Tony ElHabr: jonathan brought the jokes tonight
00:23:42 Tony ElHabr: always love that
00:23:47 Jon Harmon (jonthegeek): Thinking of %>% as "and then" is very useful. Learning to read code in a way that makes sense is super super helpful!
00:24:24 Yoni Sidi: Ironically Tony You can look at the print method for data.frames in details to see how I got around it
00:24:38 Tony ElHabr: hmm will do Yoni
00:25:08 Tyler Grant Smith: for small loops and especially if youve preallocated the for loop is probably faster
00:25:10 Daniel Chen: to fix the "[90m" stuff you need to put:
options(crayon.enabled = FALSE)
either in your renvion or on the top of the knited file
at the
00:25:25 Yoni Sidi: :+1:
00:25:36 Tony ElHabr: thx y'all. will try that out!
00:26:03 Darya Vanichkina: To evaluate the speed you can use microbenchmark::microbenchmark(<code here>) - I tend to use it when picking from stackoverflow
00:26:23 Tan Ho: ope TIL that's called a twiddle
00:26:28 Vasant M: i thought ~ is tilde
00:26:34 Jon Harmon (jonthegeek): bench::mark is another option (by Jim Hester)
00:26:38 Tan Ho: today-I-learned
00:27:03 Daniel Chen: yes it's an annononyoms fuunction
00:27:05 Vasant M: Statistician here and I call it tilde
00:27:07 Tan Ho: I always called it a lambda
00:27:09 Daniel Chen: the twittle is a shortcut for the lambda
00:27:27 Tan Ho: Hadley renames things outta nowhere for no reason
00:27:28 Tan Ho: grah
00:27:32 Jon Harmon (jonthegeek): Good to hear that, Vasant! But it isn't just Hadley!
00:27:34 Tan Ho: insert pooh meme
00:27:49 mayagans: And twittle is writing a model in twitter
00:27:51 Vasant M: Ya! I think it’s the atlantic/pacific divide
00:28:01 Tony ElHabr: zoom really missed the boat by not allowing images to be embedded in chat
00:28:08 Darya Vanichkina: .x is also hidden from the global environment (i.e. it exists but is not visible in the environment, from memory (?)
00:28:11 Jon Harmon (jonthegeek): https://hsm.stackexchange.com/questions/7999/why-do-mathematicians-call-twiddle#:~:text=Tilde%20is%20from%20Spanish%20tildar,pretty%20apt%20name%20for%20it.
00:28:30 mayagans: Im still trying to find the JB quote about loops
00:28:32 Tyler Grant Smith: i prefer giving arguments in the ... to map when possible rather than making a partial lambda. opinions?
00:28:33 Yoni Sidi: Wait … isn’t ~ $/sim$?
00:28:59 Darya Vanichkina: Tyler, can you give an example?
00:29:05 Tan Ho: > JB quote about loops
someone has to write a for loop, but it doesn't have to be you!
00:29:13 mayagans: Lol yes thank you tan so good
00:29:20 Tony ElHabr: yea tyler. ellipses as the default, and partial only minimally
00:29:57 Tan Ho: yeah, where possible I write map, map-args, function, static args, but not always possible with argument order
00:30:04 Tyler Grant Smith: so for the given example. map(1:5, my_sqrt, weight = 2)
00:30:25 Tan Ho: tbh Tyler started it
00:32:15 Kevin Kent: Yeah I like the … static arg style too.
00:32:54 Darya Vanichkina: And if we have more than one arg would it be (ex_ weight = 2, height = 3, randomarg = 4)?
00:33:03 Daniel Chen: to tyler, re ... : the ... works, but you can't map a vector of values with it (can't do: map(1:5, my_sqrt, weight = 11:15) ). that's what map2 and pmap are for.
00:33:16 Tan Ho: LIVE CODE
00:33:18 Tan Ho: YAY
00:33:26 Joe Sydlowski: The correct window setup too!
00:33:29 Tony ElHabr: insert environment pane comment
00:33:32 Tan Ho: BOOO
00:33:40 Tan Ho: BOOOOOOOOO
00:33:41 Tyler Grant Smith: yep daniel
00:34:05 Tyler Grant Smith: and Darya, yes
00:34:09 David Severski: Source goes on the right. Console on left. Change my mind. ;P
00:34:17 Tan Ho: oh nooooo
00:34:18 mayagans: Thats chaos
00:34:25 atoumi: this is absolutely the correct layout
00:34:30 mayagans: ^^^
00:34:32 Tan Ho: where's daryn I need backup
00:34:34 Darya Vanichkina: @atoumi I agree
00:34:36 mayagans: Queen has spoken
00:34:37 Kevin Kent: haha. My eyes would get crossed with source on the right
00:34:40 Tyler Grant Smith: that hurts my brain to think about
00:34:40 Daniel Chen: i have to say it's the "wrong env pane setup" when you're teaching (new learners) becuase students ask "mine don't look like that". unless you take out time from the lesson to show them how to customize it
00:34:57 Darya Vanichkina: @Daniel I actually do that at the start of every workshop
00:35:13 David Severski: Now it’s time for the RMarkdown notebook output to console/in-line debate. :D
00:35:26 Darya Vanichkina: Because that allows me to stick EVERYTHING into the bottom right pane, and have source take up the entire left of the screen
00:35:28 Tan Ho: i assume we don't need to have a lightmode darkmode fight?
00:35:39 Tony ElHabr: i think we are all on the dark side, right?
00:35:43 atoumi: Output to console and darkmode because we care about our eyes
00:35:44 Scott Nestler: That might be something we all agree on.
00:35:46 Tony ElHabr: RIGHT Y'ALL?
00:35:47 Darya Vanichkina: Also, I think it’s a discussion, not a fight …
00:35:59 Tan Ho: shush
00:36:00 David Severski: For teaching, Studio Cloud or a dockerized version is my goto setup.
00:36:08 Tan Ho: we don't pull punches with dark mode
00:36:50 David Severski: Yay! Rsthemes!
00:36:53 mayagans: Blindeddd by the lightttt
00:37:07 Darya Vanichkina: I’m Carpentries-trained/based, so I still tend to want people to walk away with something they can work with on their own data later
00:37:11 Darya Vanichkina: Although it does take time
00:37:21 Tony ElHabr: makes sense Darya
00:37:40 Daniel Chen: i run into horizontal space issues when i make the font bigger when i have to teach, but that usually ends up being workshop dependent.
00:37:43 Joe Sydlowski: I never noticed that before!
00:38:25 Darya Vanichkina: Surprisingly, I recently had a horrific workshop (geospatial python with DL, via zoom - install issues every session), and the post-workshop survey learners STILL said they wanted us to fix installs (instead of a docker-based VM solution)
00:38:43 Darya Vanichkina: [I am still setting up a viable docker/VM for next year]
00:39:11 Daniel Chen: for python, can you use a binder instance for workshops?
00:39:27 Darya Vanichkina: I needed GPUs
00:39:33 Daniel Chen: :(
00:40:00 Daniel Chen: conda environment.yml files aren't able to install the binaries?
00:40:28 Connor Krenzer: Could you use the map() function on this dataset?
00:40:29 mayagans: Your python talk is gonna get you booted from this chat (lol jk xoxo)
00:40:32 David Severski: I kinda hate the warning messages the new dplyr spits out about groups (yes, I know they’re configurable).
00:40:34 Jordan Krogmann: Great job!
00:40:37 Tony ElHabr: discussion of python install issues in an R meeting? how fitting
00:40:54 Yoni Sidi: Great job Jonathan!
00:40:54 Tan Ho: i threw the option into my rprofile once and have never had problems again
00:41:02 Darya Vanichkina: @mayagans/@Tony I think it’s more a teaching issue
00:41:08 Darya Vanichkina: And I definitely use/teach both
00:41:10 Tony ElHabr: yeah, global Rprofile option ftw
00:41:19 Tony ElHabr: don't get me wrong, i love python too
00:41:26 Tony ElHabr: i just like to joke haha
00:41:28 David Severski: I do to, but I use a lot of renv these days, so I’m always going into environments where my rprofile isn’t active
00:45:14 Vasant M: Especially when you apply a lot of different functions (models) , being explicit is better
00:45:19 Tan Ho: https://imgur.com/WdcWnKz
00:45:54 Connor Krenzer: Is there a reason why you would use map() instead of vapply()?
00:46:52 Daniel Chen: vapply lets you specify the output types and how many elements in each list element.
map only returns a list.
00:47:28 Daniel Chen: that's map vs map_*
map will return a list
00:48:25 David Severski: And the easy path to parallelization with `furrr` is so nice.
00:49:01 Darya Vanichkina: @David agreed
00:49:56 Yoni Sidi: https://github.com/hrbrmstr/freebase
00:51:02 David Severski: He really called it freebase? :P
00:51:06 Darya Vanichkina: Thanks, everyone!
00:51:08 David Severski: Bye!