15.5 Meeting Videos

15.5.1 Cohort 1

Meeting chat log
00:07:17    Jordan Krogmann:    @jon the power of the doge shirt!
00:13:59    Jordan Krogmann:    ggpairs is a great package for this
00:14:20    Conor Tompkins: I always want to tilt corr matrices a little bit
00:14:22    Asmae Toumi:    eyes emoji
00:14:36    Jon Harmon: 👀
00:26:51    Jon Harmon: https://github.com/tidymodels/workflowsets/pull/48
00:34:23    Jim Gruman: gotta step off … thank you Tony!!!

15.5.2 Cohort 2

Meeting chat log
00:08:13    Luke Shaw:  https://www.datakind.org/
00:11:46    Luke Shaw:  https://www.tidyverse.org/blog/2021/05/choose-tidymodels-adventure/
00:12:54    Luke Shaw:  https://www.youtube.com/watch?v=2OfTEakSFXQ
00:44:23    shamsuddeen:    Is ok for me too

15.5.3 Cohort 3

Meeting chat log
00:16:29    Federica Gazzelloni:    https://www.tidymodels.org/start/models/
00:17:29    Federica Gazzelloni:    https://r4ds.had.co.nz/many-models.html
00:31:46    Daniel Chen (he/him):   since there was a talk about "the whole thing" I feel like this set of examples is a good candidate for it all wihtout having to we-write large swaths of things
00:32:54    Ildiko Czeller: exactly! I feel like some of the package articles do a better job showing the "whole" thing at one place than the book
00:34:28    Daniel Chen (he/him):   the book is more for learning ml and concepts, but since tidymodels is new (to me) i'm always fighting with how I used to do things or how things are done manually
00:35:17    Daniel Chen (he/him):   I also haven't fit a predictive model since dissertation things. so I know for me personally my confusion and struggle is mainly not actually using it regularly
00:36:57    Ildiko Czeller: for me too I currently do not need predictive models for work so I need to find use cases outside work... which I haven't done yet but plan to
00:38:32    Ben:    I thought it was only me, I can relate to your struggle, its actually a thing. 
00:53:30    Ildiko Czeller: workflow_map has a default value fn = "tune_grid", which I would prefer always write explicityly. it is a sensible default but it is just weird to read the code without it for me. like we map without specifying a function
00:55:24    Daniel Chen (he/him):   +1
00:58:39    Daniel Chen (he/him):   40 minutes : |
00:59:00    Daniel Chen (he/him):   the verbose = TRUE is probably there so you know it's doing something. and didn't just stall
00:59:09    Ildiko Czeller: I wonder which of the 10 models makes it so slow. maybe neural net?
00:59:52    Daniel Chen (he/him):   3min * 12models. that's not too bad
01:00:29    Ildiko Czeller: hmm there are several models taking several minutes there. interesting. it would probably worth it to set up computing in the cloud on a more powerful computer if it is important.
01:01:18    Ildiko Czeller: yeah it is not too bad to wait once, but if you realise in the end you made a mistake...
01:05:26    Daniel Chen (he/him):   that was a good review of the most confusing chapters
01:11:28    Daniel Chen (he/him):   i have conference stuff next 2 weeks so i'm +1 for break

15.5.4 Cohort 4

Meeting chat log
00:23:21    Federica Gazzelloni:    https://www.tidymodels.org/find/#search-parsnip-models
00:23:27    Federica Gazzelloni:    https://baguette.tidymodels.org/
00:23:57    Federica Gazzelloni:    mars Multivariate adaptive regression spline
00:25:21    Ryan Metcalf:   [SubModel Optimization](https://www.tmwr.org/grid-search.html#submodel-trick)
00:39:01    Steve C:    https://cran.r-project.org/web/packages/Cubist/vignettes/cubist.html
00:39:05    Steve C:    cubist
00:57:32    Brandon Hurr:   https://stats.oarc.ucla.edu/other/mult-pkg/whatstat/
01:10:12    Isabella Velásquez: Thank you!!