For all interested in #sampling #survey and #rstats I would like to endorse our new package: {nonprobsvy}. It implements state-of-the-art methods for #nonprobability samples published in top #stats journals (and many more!). Feel free to check it out! Development version can be found at https://github.com/ncn-foreigners/nonprobsvy
#sampling #survey #rstats #nonprobability #stats
Is there a classical regression model where, for π=1,β¦,π,
πΈ(πα΅’) = π πα΅’
with π a known constant, and
πα΅’=(exp πα΅’ Ξ²) / (ββ±Ό exp πβ±Ό Ξ²)
Thus πα΅’ β (0,1) and β πα΅’ = 1.
Note that this is *not* a multinomial logistic regression. There is a single vector Ξ² to estimate. It should be estimated from a single set of observations πβ,β¦,πβ (and the covariates πβ,β¦, πβ).
#rstats #glm #statisticalmodel #statistics
It's back to school so how about learning how to write #reproducible #datascience pipelines with #RStats ?
My book "Building reproducible analytical pipelines with #rstats" can help you with that!
Buy a drm-free epub or pdf @leanpub
https://leanpub.com/raps-with-r/
Physical copies are available https://amazon.com/dp/B0C87H6MGF (check your local market place)
Read for free forever: https://raps-with-r.dev
#reproducible #datascience #rstats
Our main activities are relatively similar between countries, with some differences depending on economic development (based on this week #TidyTuesday )
#RStats code: https://github.com/BjnNowak/TidyTuesday/blob/main/SC_HumanDat.R
This little side quest into unnamed data.frames was because {plotly} prepares data in unnamed data.frames in order to get {jsonlite} to output data in a particular JSON array format, rather than a JSON object format.
TIL.
And {dplyr} won't let you select columns which don't have names.
Even if you're trying to select by numeric index
{tibble} will let colnames be set to NULL, but then has issues printing the tibble to console.
TIL: It is possible to have a data.frame without column names.
Not sure why you'd want to though...
@milesmcbain I'd appreciate only recycling for length = 1.
All other cases you have to do an explicit rep() / rep_along()
I think I just generated my favorite {codename} project name ever: "mulberry monkey" π«π #rstats
Vector index recycling is convenient but it would be really great if #rstats core could sort it so that it plays a sound effect of some chimes tinkling when itβs happening so I know to check I want that magic right now. Next R sprint maybe? β¨π§ββοΈπͺ
π’ R Consortium Pilot 3 Mileston π’
Big news! R Submissions Working Group successfully submitted the pilot 3 R-based package to FDA eCTD on August 28, 2023. This is the first public R-centric FDA submission!
#rstats #fda #datascience #opensource
A map showing median household income of New Jersey, but this time by population (the previous one was census tract).
Higher spikes indicate greater population, colours indicate median income.
Ok, #rstats markdown avengers, assemble
Is there a way to extract specific options (namely fig.cap, tab.cap) from all the chunks in an RMarkdown document, export those to a text file, and then , assuming those options get edited externally by colleagues, read them back in again.
Trying to avoid tedious, tortuous and error prone copypasta between Word's track changes and my Rmd file..
Is this possible? If so, what's the best approach?
Feeling so, so stoked to see #rstats friends at posit::conf(2023) next week.π
Find me for a limited edition raspberry variant of the stacks hex.π
Catching up on last week's #TidyTuesday, and keeping it simple with some line charts! This plot uses two of my favourite #ggplot2 extension packages:
πΊοΈ {geofacet} for visualising data across different geographic regions
π {gghighlight} for highlighting a subset of values in a chart, here showing each state in relation to others
Code: https://github.com/nrennie/tidytuesday/tree/main/2023/2023-09-05
#tidytuesday #ggplot2 #rstats #dataviz #r4ds
So very very excited for posit::conf in a week or so! Can't wait to meet up with so many of my favourite #rstats folks, and meeting loads of new ones too π
#geobr and #censob by @UrbanDemog are terrific #rstats packages! Saw a fun vignette and thought I try making it 3D because why not? Gratuitous, possibly. Pretty, I think so.
Here's the proportion of households in Brasil connected to a sewage network in 2010.