With the latest r-universe that provides CSV, JSON, XLS, etc. for datasets in #rstats packages, you don't even need to install the package to access the data. You can read them directly from the r-universe site. It's very convenient for teaching!
It's powered by #webr as part of the r-universe high-performance node.
The {pracpac} #rstats 📦 aims "to provide a usethis-like interface to create Docker images from R packages under development. . . .
"[It] can be used to containerize any R package and deploy with other domain-specific non-R tools, Shiny applications, or entire data analysis pipelines." https://signaturescience.github.io/pracpac/
By VP Nagraj and Stephen Turner at Signature Science @rstats
You got my interest peaked (or whatever). I like stats, but in the moment and "seat of the pants" (as tree-top flyers say). And I am pretty good at it, like 99% on. But I need this. I need to see how to fit R in.
If you ever need to create fake data (for class, simulations, etc.), the {charlatan} package makes it really easy to create dummy data for names, phone numbers, emails, numbers, genes, & more! #RStats
Realized I never wrote an #introduction after rejoining.
I'm a hospital data analyst (data scientist if you're nasty) who loves :rstats: #Rstats and toying with #QuartoPub to automate more and more of my work.
Started life as a journalist, now that punk rock ethos pours out as a passion for health equity.
Does anyone know if you can make RStudio change the spellcheck language per document? Writing a multilingual blog can be annoying since I need to change the language of the project and then reopen the project.
Another little #rstats dependency thought experiment for you.
Setup: Let’s assume that coding style has a significant impact on robustness to dependency failure. I’ll argue this is obviously true because DRY code (code with less duplication) is easier to refactor since a) there’s just less of it, and b) it tends to be more abstract with function or object interfaces acting like dependency effect quarantine zones…
@milesmcbain it’s really unclear to me why this has become some huge conversation. From a practical stand point, nothing has been less of an issue for me in over a decade of R programming, and I’m on the side of running code unattended on servers “in production”, which arguably is harder to deal with than purely reproducible research.
I'm super excited that later today at 10:30 Eastern Time I'll be speaking at the Data+Donuts series at Harvard Kennedy School about the Public Health Disparities Geocoding Project 2.0.
Among other things, I'll talk a little bit about the role the #tidycensus package and GIS software in R like #sf can serve in analyzing data to advance health justice.
@debruine it's really handy! We included a commented groundhog version of package loading commands in our latest paper which can just be uncommented for easy computational reproducibility once the package versions we used become EOL.