Data Science. TileDB. Open Source. Quant Research. R. C++. Debian. Linux. Adjunct Clinical Professor, University of Illinois. Lots of coffee. And some running.
There are lots of nice latex styles that are wrapped. See #Rstats packages
- `rticles` with many journal styles
- `binb` with three different beamer styles
- `tint` with a spin on the Tufte style
and more. For `rticles` I put up a gallery here
github.com/eddelbuettel/rtic…
When I first wrote 'crp' it started web scraping. Having svn or git commits is so much more powerful -- but then we all also already have access to r-devel changes.
The next move, if any, should come from any combo of R Core and CRAN to make 'policy' changes public for us.
New r2u demo showing easy, reliable installation of three CRAN #RStats packages onto @Ubuntu 20.04 which had otherwise frustrated a SO user. With gif :)
Answer: error in library.dynam(lib, package, package.lib) shared ovject <library_name>.so not found stackoverflow.com/a/73351708…
The CRAN Task View on High-Performance Computing has some recommendations on 'big data' with #RStats
Seehttps://CRAN.R-project.org/view=HighPerformanceComputing and/or its markdown file github.com/cran-task-views/H…
Make it a factor! #Rstats
> vec <- c("red", "blue", "red")
> table(vec)
vec
blue red
1 2
> fvec <- factor(vec, levels=c("red", "blue", "green"))
> table(fvec)
fvec
red blue green
2 1 0
>
Delete all those tweets ASAP.
You never get an academic (or industry) job using `rm(list=ls())` as all your colleagues will be terrified expecting their work to get pulverized too.
Seriously, it's a rather bad habit. Write #Rstats scripts and work in clean, fresh shells.
Neat story if you've ever listened to the phenomenal album "My Life in the Bush of Ghosts". And if you haven't yet what the heck did you do the last 40 years?
‘Better late than never’: how @brianeno and @DBtodomundo finally laid a musical ghost to rest theguardian.com/music/2022/a…
Old school:
- install into, say, /opt/R-4.2.1, /opt/R-4.1.3, /opt/R-devel and as many others as you want
- call respective version, either by adjusting $PATH or explicitly with full path
- once you 'run' that main R process, it knows its directies and libs and "just" works
Yes I enjoyed using it during the few years our campus had a it as a pilot. Small nuisances (eg new dotfiles and hence git setup for each project) but overall good. We will now pivot back to RStudio Server on our own server for all students. Likely easier, and no extra fees.
There are several CRAN packages allowing Julia to be called from R (JuliaCall, JuliaConnectoR) and there papers published on them. Plus there is the (excellent, though now dated) discussion in Chambers 2016 book "Extending R". Can you dig in an provide a review? #rstats
Well we happily do 'the opposite': packages with _full and complete dependencies_ that are moreover shipped as _binaries_ so you don't even need to build. See eddelbuettel.github.io/r2u/ for @ubuntu 20.04 and 22.04, similar repos exists for Fedora and OpenSUSE. #rstats#r2u