Data Science. TileDB. Open Source. Quant Research. R. C++. Debian. Linux. Adjunct Clinical Professor, University of Illinois. Lots of coffee. And some running.
Chicago, IL, USA
Joined March 2007
- Tweets 5,361
- Following 350
- Followers 10,149
- Likes 7,646
Help welcome: @lenovo X1 unstable with @Ubuntu 21.04 (after many happy earlier releases) and locks under @Gnome. Question bounty expiring tomorrow.
Ubuntu 21.04 locks hard with basic GUI redrawing using standard intel graphics askubuntu.com/q/1346317/1771
2
1
Interested in #LiDAR data management for serverless access from #rstats, #python, #sql (and more) interfaces binding to the (open source) core library? And/or how @tiledb also adds a cloud management and sharing layer? So come to next week's webinar!
Join us for a live workshop on how a universal #datamanagement system based on multi-dimensional arrays can be a game-changer for managing, analyzing, and sharing massive #LiDAR data. Ask Stavros, the original creator & CEO of TileDB, your hardest Qs! hubs.ly/H0QBS8C0
1
5
12
RcppGSL 0.3.9 on CRAN: Some Policy, UCRT Build
Easier GNU GSL use from R
dirk.eddelbuettel.com/blog/2…
#rcpp #rstats
/cc @opencpu
3
You can always take advantage of binaries from a LTS release on a newer releases---so you have a choice of both RSPM (per Grant) or BSPM (my pref, using the PPA) for pre-made binaries making installation faster and easier.
1
1
Well in a narrow sense *every one* of the 4000+ compiled packages on CRAN can and many do so as they all (can) have Makevars dot win and configure dot win, and many use it to good effect. But you probably want to talk to @opencpu...
1
3
Dirk Eddelbuettel retweeted
Thanks @ajggeoger & @map_andrew for having us on the @eoscenefrom podcast, where we discussed how a universal database based on dense & sparse multi-dimensional arrays can unify all geospatial data. scenefromabove.podbean.com/e…
1
5
7
Dirk Eddelbuettel retweeted
@tiledb is looking very promising for our #rspatial shiny apps.
Looking forward to giving it a proper kicking & testing its limits!
#rstats
We're always looking for ways to make our #shiny apps more responsive. And it always comes down how to store and query data.
We've given @tiledb a run and the results are very promising
spoiler: ~0.4 secs to query & return the data from S3
resources.symbolix.com.au/20…
#rstats
4
4
And here is a thread about it from the @nytimes:
a team of one hundred to support the coronavirus case data collection and analysis. And, as of today, a truly well deserved Pulitzer Prize for Public Service!! Sincere congratulations!
You may have seen The NYT’s coronavirus case map. It's part of a sprawling data effort that involved over 100 journalists. We learned today our work was part of an entry that received the Pulitzer Prize for Public Service. Here’s how the project came to be.
3
Dirk Eddelbuettel retweeted
Thanks to @eddelbuettel for showing us the ropes and developing the #rstats {tileDB} library.
You can get hands-on experiece at his useR tutorial -
Come to our @tiledb #rstats tutorial where @aaronwolen and I may show how to put a 194 million rows "flights" csv file (85 gb uncompr.) into a single TileDB (sparse) array (indexed by flight date, carrier, origin and destination) you can read / write to / from S3 / GCS / Azure.
2
3
Dirk Eddelbuettel retweeted
We're always looking for ways to make our #shiny apps more responsive. And it always comes down how to store and query data.
We've given @tiledb a run and the results are very promising
spoiler: ~0.4 secs to query & return the data from S3
resources.symbolix.com.au/20…
#rstats
1
9
1
26
And as nobody reads blogs anymore, here is the direct video link with @VincentAB, @grant_mcdermott and myself in supporting roles -- and starring #Emacs, #RStats and #byobu in the lead. A thrilling nail biter about collaboration in them interweb clouds:
youtube.com/JX9m5B0BL7s
1
6
21
Yes, @RStudio Cloud is good -- I taught my data science programming class (cf stat447.com) using it three times now. Essentially zero setup issues. At U of Illinois we have a campus-wide contract, and students pay around $5/month. #rstats
1
6
R^4 #33 and T^4 Video #8: Collaborative Editing and Execution via Byobu
Share a session with collaborators to jointly edit and run code, including #RStats
dirk.eddelbuettel.com/blog/2…
With @grant_mcdermott and @VincentAB
3
8
Public data, public health -- so below is some public code too. The @rdatatable and #ggplot2 combination makes it a breeze. #rstats for the win
2
6
1
30
Fifteen years ago, I gave the talk 'Use R! in fifteen different ways: A survey of R front-ends in Quantian' at useR! 2006 showing what I put into the Quantian cdrom/dvd.
Slides with screenshots (for Linux, no Win or Mac) at
dirk.eddelbuettel.com/papers…
#rstats
1
9
Yes, traceable from tar.gz release and git/svn. We have
- the Windows GUI by B. Ripley + G. Mazarotto
- Unix Gnome-2 GUI (which we wrapped as Debian binary too)
- and maxOS / osX always had something.
See video of @RogerBivand at celebRation last year for windows.
#rstats
2
2
9
Inflated count that does not correct for 'deceased' packages. The official #RStats package count is (currently) this:
2
7
Yes. But I showed Docker here because it makes the example self-contained. I run the same setup (which maybe I should document again outside of this Dockerfile) on another machine and it just *rocks* so hard that the usual `update.packages()` then pulls _binaries_. #rstats
1
1