Data Science. TileDB. Open Source. Quant Research. R. C++. Debian. Linux. Adjunct Clinical Professor, University of Illinois. Lots of coffee. And some running.
I am looking forward to talking about #RStats, #Rcpp and #ML at the 6th Europeans #COST Conference on #AI in Industry and Finance organized by four @ZHAW departments. All happening tomorrow, see more at
zhaw.ch/en/engineering/insti…
Tomorrow, we're examining all things TileDB Embedded, our open-source storage engine that universally models any data, whether it's MBs on your laptop or PBs on cloud object storage. Join us at 10am EDT and get your questions answered live. Register at hubs.la/H0WRYWy0.
Confession: Signed up for @letterboxd as @j_v_66 mentioned it as a) good and b) possibly underused.
So far, so good. Mostly playing review ping pong with Jan (who surely returns more balls to many others).
But then ... this zinger. Putdown of all putdowns? Hall of fame.
Our excess-deaths model, launched this week, is an exceptional piece of work by @Sondreus and @martgnz. The central estimate is that 15.2m deaths have been caused by covid-19, more than three times the official toll of 4.6m
What is the pandemic's true death toll? After months of work, we are launching our daily updating estimate of excess deaths around the world. With official deaths at 4.5m, I estimate the true tally to be between 9.3m and 18.1m, as of today
Open Source is the best. At one point #simdjson made me curious so I built an #rstats package. In walks @knapply_ extending it to vastly outperform all R alternatives. It rests ... til @lemire and Fred Boyer take a spin and update it again. Just wow.
New release very soon.
Join us for a live deep-dive on the internal mechanics of TileDB. Learn about the open-source TileDB Embedded array storage engine and how it serves as the foundation for the first universal database, TileDB Cloud. Register for Sept. 9 at 10am EDT. hubs.la/H0VWxg10
The end of an era, and, if we may add, good riddance.
A bunch of mailing list (and, e.g., @stackoverflow) answers for #rstats will be invalid but that is the sweet, sweet taste of slow moving progress (in, as usual, r-devel aka R 4.2.0 to be by next spring).
The older I get, and the more I use #RStats—20 years now—the more I come around to reducing package dependency. There is SO MUCH in base R that just gets the job done, almost always more quickly, & it makes your code more robust. Code I’ve written 15 years ago still just works.
Any other family with "friendly" competitions over the @nytimes (mini)-crossword? My college-age daughters are _shredding_ me. Could it have to do with using phones (them) versus the browser (me)? Any studies out here? Can I cheat otherwise? Asking for a "friend".
Data Science Programming Methods is back as STAT 447 for Fall 2021, and just started. Expanded shell programming, two guest lectures, the usual mix of git, md, sql, and #RStats will make for am exciting term, for more see stat447.com@Illinois_Alma@IllinoisStat
BTW @Enchufa2 and I have a paper at arXiv on this; it is (as much as it pains me to say this ... ) even more comprehensive on Fedora and OpenSUSE. We need a .deb based volunteer effort to catch. If I had more time...
arxiv.org/abs/2103.08069