Data Science. TileDB. Open Source. Quant Research. R. C++. Debian. Linux. Adjunct Clinical Professor, University of Illinois. Lots of coffee. And some running.
Congrats to @dselivanov_ et al for RestRserve release 1.0: an awesome package scaling #Rstats to thousands of concurrent REST requests per second (think `plumbr` on steroids). My 15s of fame here are from having suggested use of Rserve years ago...
restrserve.org/
Thanks again to @NISS_DataSci for inviting me to give a virtual tutorial on #Rcpp for extending #RStats with @isoCpp. Slides are (as usual) on my talks page at
dirk.eddelbuettel.com/talks.…; the video recording should be(come) available for NISS affiliates.
Join us on Friday, March 25 for the Rcpp (R and C++) virtual tutorial by Dr.Dirk Eddelbuettel, Principal Software Engineer at @tiledb and adjunct Clinical Professor at @IllinoisStat.
Moderator: Dr. Kevin Lee (@WesternMichU)
Maybe 'Library of Statistical Techniques'?
It accepts posts via PRs and has a fair number of #Python and #Rstats entries for the same estimation, modeling, data analysis, ... tasks
lost-stats.github.io/
h/t @grant_mcdermott
Endorse! This sounds a lot like the first four weeks of my STAT 447 (cf stat447.com) course! If someone wanted get in touch to offer this as a day-long workshop my arm could likely be twisted.
The rest of Mike's thread is good too, as usual 😉
I'm not yet qualified to give advice on the job market, but if you're an academic hoping to hit the ground running as a data scientist in industry, here are some potentially useful tips:
‣ Round out your programming skills–be okay or comfortable with git, bash, SQL.
Please help test #RStats 4.2.0
I added a first alpha release to @debian unstable: `apt install -t unstable r-base-core` will get it. I also updated the #RockerProject r-base container with tag 4.2.0. I plan to update them every few days until release on April 22. Please test.
Not really a conda user myself but I recently verified that using `mamba` via the condaforge/mambaforge container allowed me to run
mamba install -y r-$NameOfCRANPkgHrere
without issues a given #RStats package and all its dependencies. Maybe that will work?
At work we've decided to combine our favourite C++ features that enable robust, reliable, and repeatable code:
Thrand().
It is a combination of threads and rand().
I plan to propose it for C++26, although maybe it is important enough to push for a special inclusion into C++23.
Yep. Also used for the fastest (if obscure) moving average over a (time) series which is why it always was so annoying the function got stomped upon. Namespaces are precious, and warnings matter.
To clarify: while there is a Pusbullet browser plugin, the R package (like other API clients) works with the backend and does not involve the browser. So it works headless (say, at AWS), in scripts, via cron, ... you name it. But powerful and flexible for alerts and result.
The #Rcpp webinar / talk is happening in three days from now, and registration is still open (and inexpensive overall, cheaper still for students and free for those at NISS partner institutions). #rstats
Join us on Friday, March 25 for the Rcpp (R and C++) virtual tutorial by Dr.Dirk Eddelbuettel, Principal Software Engineer at @tiledb and adjunct Clinical Professor at @IllinoisStat.
Moderator: Dr. Kevin Lee (@WesternMichU)
9 years. Well, as the characters of my favorite TV series “The Wire” used to say: “You only do two days. That's the day you go in and the day you come out"
I even had a T-shirt with this slogan, but the prison authorities confiscated it, considering the print extremist.
The #Rcpp website has been refreshed, but still has quick intros, pointers to the ten vignettes, the gallery, 2500+ #Rstats packages using #Rcpp, and the book.
At rcpp.org (and yes, I should order a https cert)
Thanks all: I am (of course) aware of `ymd()`, the point of the tweet was that `anydate()` avoids all this automagically.
Plus, as @shabbychef reminded us, base #Rstats was there first (for this value) as `as.Date("2022-03-21") + 7` also works. Peace out.
I also use it but it a) supports fewer formats that anydate(), b) does not deal with as many input types and c) still insists on the re-repeating for 1e5s time what "origin" is (as if there ever was a doubt) so anydate() wins.
Apart from that it's great and zero-dependency.
> lubridate::dmy("2022-03-21") + lubridate::days(7)
[1] NA
Warning message:
All formats failed to parse. No formats found.
>
> anytime::anydate("2022-03-21") + 7
[1] "2022-03-28"
>
Always happy to help another #rstats user 😉
𝚍𝚖𝚢("𝟸𝟶𝟸𝟸-𝟶𝟹-𝟸𝟷") + 𝚍𝚊𝚢𝚜(𝟽)
That's right, for folks planning to submit your talk for consideration at 𝚛𝚜𝚝𝚞𝚍𝚒𝚘::𝚌𝚘𝚗𝚏(𝟸𝟶𝟸𝟸), we’re extending the deadline to March 28.
#rstudioconf#rstudioconf2022
Join us on Friday, March 25 for the Rcpp (R and C++) virtual tutorial by Dr.Dirk Eddelbuettel, Principal Software Engineer at @tiledb and adjunct Clinical Professor at @IllinoisStat.
Moderator: Dr. Kevin Lee (@WesternMichU)
If you want to learn about #Rcpp for C++ and #Rstats integration, I will be giving a one-hour introduction on March 25 thanks to a kind invitation by @NISS_DataSci via Kevin Lee. Registration should be opening soon, more details at the link below.
niss.org/events/rcpp-r-and-c…
If you want to learn about #Rcpp for C++ and #Rstats integration, I will be giving a one-hour introduction on March 25 thanks to a kind invitation by @NISS_DataSci via Kevin Lee. Registration should be opening soon, more details at the link below.
niss.org/events/rcpp-r-and-c…