Helped someone debug some tidyverse data processing issues. It turns out "NA" was a legitimate code used in their data and readr by default interprets it as NA, not a string. Careful folks! #rstats
Edit: for anyone who doesn't know, read_csv() has an na parameter. The default is na = c("", "NA"). Setting it to na = "" fixed the issue.
like most human skilled activities, both should be done by people who are knowledgeable (often not the same person), whatever tool they use.
But some tool are usually better than other.
After 30+ years in the profession, I can tell you that using a programming language like #rstats provides at least a much better and reproducible workflow than any other C&P tool that you can think of, and if you think otherwise it's only because you're like a stakeholder in SAS 😉
@odr_k4tana@sharoz
in some ways the raise of the #tidyverse sect has funnelled as side effect a general less knowledge of #rstats When I see people call readr::read_rds instead of readRDS I just despair
What does the network of people with common package contributions look like?
And much more!
There are also interactive {gt} tables so you can browse contribution and package statistics; and I’ve shared the data so you can explore your own questions. Enjoy!
On a totally unrelated note, I still don't understand who is out there using reticulate frictionlessly. Every single time I try to set anything up I get errors upon errors.
The first one talks about non-compartmental analysis and has some very simple R code, the second one dives a little deeper and talks about compartmental models - which often require you to use an ODE solver in the model - in Stan and R.
I'm still learning pharmacokinetics so they're a little rough around the edges
#rstats package {crul} 1.4 is on CRAN. It comes with many new features, I'm really excited about retry on Async requests. Thanks again @sckottie for the all the work on this package.
The goal was to learn about applying splines to a circle for the polygon chapter, but splines are too hard for my brain right now, so this resulted instead. Forever searching for less computationally heavy ways to make space 🌃 #rtistry in #rstats w/ geom_polygon() & geom_point()
I'd totally forgotten about this quirk of #rstats functions: arguments are not evaulated until they are used, so if argument b defaults to the value of argument a, you need to use argument b in the code before you make any changes to a (or of course don't change a)
((I spent half an hour debugging something due to this today))
x <- function(a, b = a) {
a = 1
return(b) # first use of b sets the value b = a = 1
}
x(2) # returns 1
y <- function(a, b = a) {
b # sets the value of b = a = 2
a = 1
return(b)
}
Die Slides, Notizen, Code und Vertiefungshinweise für meinen Vortrag "Legal Data Science: Der moderne Weg zur Wahrheit" bei der Digitalen Richterschaft sind jetzt online!
“It’s a bit like seeing a rabbit shape in the clouds and then testing whether all clouds look like rabbits… using the same cloud. I hope you appreciate that you’re going to need some new clouds to test your theory.
Any datapoint you use to inspire a theory or question can’t be used to test that same theory.”
Things I do while waiting for huge SAS files to process for work 😴 AKA: "What is this sorcery I've managed to conjure up with geom_segment() 😭?" #rtistry#rstats
When you struggle with theme()-arguments in {ggplot2}, you may like {ggeasy}, offering shortcuts to plot customization https://github.com/jonocarroll/ggeasy (worth visiting just bc of the hilarious cartoon)
I am skeptical when a package promises to make plotting easier because you have to remember the arguments you are looking for in either case. Here, however, I can see the added value for some arguments #rstats#dataviz@rstats
Trying something different. Again. A map of Japan's shinkansen network as of 2015. If you know an updated source, please let me know. Like the Interstate 5 map, the camera angle and hovering lines may give the impression that they are off; they are not. 😅
I asked Bard how to read in a .mseed file into #rstats
If the seismology package existed and worked in this way I would be interested. But in terms of getting up and running with this specialist binary file, I would have been better (had I not known my options before hand) to use a more traditional source that restricted itself to indexing websites rather than making up convincingly communicated answers.