mccarthy-m-g / tidytales Goto Github PK

Post title: Unlearning statistics

Intro idea: I never learned anything about statistics until the first year of undergraduate psychology. And I was surprised I needed to learn about statistics, if I’m being honest (I’m one of many who went into psychology with the goal of helping people through clinical counselling). My first impression was that statistics was just math. I was good at math so I wasn’t too concerned. However, what I encountered instead was some strange discipline with all these tests with different names, supposedly important numbers called p-values, and a framework these tied into called NHST…

A lot of claims about how the world works capitalize on phenomena that have nothing to do with how the world (e.g., http://deevybee.blogspot.com/2013/06/interpreting-unexpected-significant.html)

Mainly as something of a reference list for papers that try to improve statistical thinking:

• https://digest.bps.org.uk/2019/03/12/psychologists-love-to-report-marginally-significant-results-according-to-a-new-analysis/amp/
• https://journals.sagepub.com/doi/10.1177/2515245918771329
• The null ritual https://pure.mpg.de/rest/items/item_2101291/component/file_3080636/content
• https://www.ime.usp.br/~abe/lista/pdfVZSdqrML7E.pdf
• https://www.rips-irsp.com/articles/10.5334/irsp.198/
• https://www.rips-irsp.com/articles/10.5334/irsp.82/
• https://pubmed.ncbi.nlm.nih.gov/19017432/
• https://journals.sagepub.com/doi/full/10.1177/2515245919847202
• https://psycnet.apa.org/doiLanding?doi=10.1037%2F0033-2909.97.1.129
• https://projecteuclid.org/journals/statistical-science/volume-16/issue-3/Statistical-Modeling--The-Two-Cultures-with-comments-and-a/10.1214/ss/1009213726.full
• https://journals.sagepub.com/doi/10.1177/0959354311429854
• https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4828473/
• https://journals.sagepub.com/doi/full/10.1177/2515245918770963
• https://www.tandfonline.com/doi/full/10.1080/00031305.2019.1583913
• https://projecteuclid.org/journals/annals-of-applied-statistics/volume-15/issue-3/The-ASA-presidents-task-force-statement-on-statistical-significance-and/10.1214/21-AOAS1501.full
• https://projecteuclid.org/journals/annals-of-statistics/volume-33/issue-1/Analysis-of-variancewhy-it-is-more-important-than-ever/10.1214/009053604000001048.full
• https://pages.ucsd.edu/~cmckenzie/Simonsohnetal2014JEPGeneral.pdf
• https://journals.sagepub.com/doi/10.1111/j.1467-9280.1997.tb00536.x

The hacks for shared axis labels in patchwork presented in this snippet should no longer be necessary thanks to this PR: thomasp85/patchwork#337. The post can be updated with an example of the new axis collection functionality at the top.

Using data from tidytuesdayR::tt_load("2021-06-01")

From Tidy Tuesday https://github.com/rfordatascience/tidytuesday/blob/master/README.md

Some of the stuff mentioned here https://fosstodon.org/@mccarthymg/110667330256963483, plus Quarto and things like rendered Word doc comparisons when collaborating with non-technical dinosaurs.

Testing Utterances comments. The page this was for is no longer on the site.

Using data from tidytuesdayR::tt_load("2019-04-23")

Using data from tidytuesdayR::tt_load("2020-08-11")

https://github.com/rfordatascience/tidytuesday/tree/master/data/2022/2022-01-18

Maybe try out the ordinal package with rating as DV?

Quarto now supports custom giscus themes. Just need to wait for v1.3 to release to use a custom theme.

Using data from https://github.com/jbryer/brickset

And some inspiration: https://public.tableau.com/app/profile/mateusz.karmalski/viz/LegoTechnicironviz/start

The Star Wars or Lord of the Rings set themes might be fun to visualize

Exploring special cases in statistics (e.g., how everything is a linear model)

Add a new section to the website for posts of code snippets showing how to do something without as much exposition as a typical post.

example: https://sciencificity-blog.netlify.app/code-snippets.html

Using data from tidytuesdayR::tt_load("2021-05-25")

From Tidy Tuesday https://github.com/rfordatascience/tidytuesday/tree/master/data/2022/2022-09-06

Could maybe use that lego plotting R package

Particularly for the title block on the home page, it feels a bit too heavy. Maybe 500 weight could work better

Edit: it feels fine on desktop. Just on mobile it feels heavy.

https://github.com/rfordatascience/tidytuesday/blob/master/data/2020/2020-05-05/readme.md

Accidentally removed when switching to Quarto. See https://quarto.org/docs/websites/website-listings.html#feeds

Using data from tidytuesdayR::tt_load("2020-03-03")

Apparently Quarto doesn't just include images declared in the social metadata params into the site, so they need to be explicitly added using the resources param

https://github.com/rfordatascience/tidytuesday/tree/master/data/2022/2022-12-27

See https://www.pipinghotdata.com/posts/2021-04-01-deploy-previews-with-netlifly/

Finish up draft and publish

https://github.com/rfordatascience/tidytuesday/tree/master/data/2022/2022-10-18

To the new one

Using data from tidytuesdayR::tt_load("2019-08-06")

From Tidy Tuesday https://github.com/rfordatascience/tidytuesday/tree/master/data/2021/2021-08-17

can use trekverse for plot theming

Using data from tidytuesdayR::tt_load("2021-04-20")

Unit testing {targets} pipelines with {testthat}.

Purpose of the post is self-explanatory based on the title. I might be able to use Tidy Tuesday data if there are some messy examples that require wrangling.

Probably best to go with a simple pipeline with data import, and a few data wrangling steps. Then show how testthat can be used to test that the data meets certain criteria after being wrangled (e.g., values within certain range, etc.).

Using data from tidytuesdayR::tt_load("2021-03-16").

Using data from tidytuesdayR::tt_load("2019-07-30")