Sorry, I meant Rmakdown on top of Databricks has NOT worked well. Dang autocorrect. π
Sorry, I meant Rmakdown on top of Databricks has NOT worked well. Dang autocorrect. π
Rmarkdown on top of Databricks has worked well, but if you're unable to form a direct connection to a serverless SQL warehouse, I'd roll back to using sparklyr and run that SQL within a code chunk of your remote Rmarkdown (not part of the docs directly, below), but I'd reference the sparklyr docs:
I'd recommend using {brickster} which allows you to connect to a serverless SQL warehouse, bypassing the need to establish a connection to a Databricks Spark cluster altogether. It avoids the {odbc} dependency and facilitates a {DBI} + {dbplyr} backend: databrickslabs.github.io/brickster/ar...
Unfortunately yes I do. Is the end goal to read / write to a Unity Catalog table? Or are you trying to do more modeling using Spark (and therefore need to connect to a Databricks cluster)? If you're just wanting to connect to delta tables for read/write capabilities,
I love this idea: duckdb.org/community_ex.... It translates dplyr syntax _inside_ of duckdb so you can mix it with regular SQL
Awesome opportunity to contribute to open source ML tooling at @posit.co (f.k.a. RStudio), a Public Benefit Corp AND Certified B Corporation. Posit sells enterprise software that allows them to hire FTEs that contribute to open source projects, and now their OSS ML team is hiring for a Python dev! π―
Found this on LinkedIn but truer words never spoken. Stop the Python project "dependency hell" and make the switch to Python's {uv} framework. Here's a great YouTube video that really resonated with me being an R first dev... From Posit's Julia Silge: youtu.be/pMVYl9fx1EE
I love these slides... Thank you!
Hoping the cut wasn't accompanied by an expensive ER bill! π
I feel the same way Libby... I didn't even realize there was SO MUCH #rstats on Twitter until it was too late. Yes, I had an account for like a decade, but two months after I decided to become more active on Twitter, everyone left (including me). π Terrible timing.
EmbeddingGemma looks awesome π
πππ
SoCal RUG is back! @emilhvitfeldt.bsky.social will be demoing Orbital, a framework to convert your fit model workflows from scikit-learn (Python) or tidymodels (R) to native SQL for in-database prediction! πΈ Please RSVP for this hybrid in-person + YouTube Live event: meetu.ps/e/Py49Y/v02v...
Fascinating stem cell research from the University of Zurich!
www.sciencedaily.com/releases/202...
This book is ultra dry, but amazing. I read this when it was first published but your reminder is encouraging me to re-read it. π
In 2022, inspectors described a βcascade of failureβ at one foreign factory.
The FDA barred it from shipping drugs to the U.S. β then excused more than 20 drugs from the ban, despite risks to consumers.
Truer words never existed π
Even on Mac, eh? π Adding this to my Monday morning list. TY!
Quarto 1.7 is here! π
Now you can easily add light/dark mode to your projects with brand files and control which computational outputs appear in each mode with the new renderings option.
See it in action on quarto.org and learn more: quarto.org/docs/blog/po...
#Quarto #QuartoPub
We are in the process of getting some servers deployed with GPUs (outside of Databricks) and I'm eager to test Stan (and XGBoost!) on these servers. Hopefully I'll have something more informative to report in a few months.
I deal with this often too. Good solution here!
SoCal RUG is hosting Karandeep Singh on April 24! He released Tidier.jl in Jan 2023, basically R's {dplyr} for Julia, and with the help of OSS contributors, there are now 10+ packages in Julia that implement {tidyverse} packages! π₯ Please RSVP for in-person event here: www.meetup.com/socal-rug/ev...
RSVP (for in-person only) is now live: meetu.ps/e/P1hJB/v02v...
Today I learned about k3s... Cool stuff!
Throw a "fuck" or "fucking" into your search to bypass the AI Overview
I've just found the best hack. Swear in your search request. It's a great way to bypass the AI summary and make your search history look like you're chronically pissed off at everything!
I presented these slides last night to Tech by the Beach (@nicmoe.bsky.social π) and CSULB's MSIS graduate program. It was fun to highlight workflows and benefits with the Positron IDE, Quarto, Arrow, and more! This is not 100% mobile-optimized. #pystats #rstats #datascience rrr.is/BigDataTinyL...
Just rented the independent #documentary Pistachio Wars. If you're interested in some of the history of #California #water rights, this is a phenomenal watch that exposes the greed behind the private hands that control so much of California's water for profit. gathr.com/vod/537d592b