Still messing around with using OpenAI LLMs in RStudio for writing assistance tools. Getting some decent results from a shiny app that I'm building for grammar checking. This package is experimental and could change at any moment as I figure out what I want to do with it, but it's still fun to make hex logos and pretty things up using pkgdown. (posting to see if the logo shows up here :) )
The #rstats tidymodels team recently converted our website to use #QuartoPub, allowing us to more easily maintain and contribute content.
There are a few bonus features that came along with this conversion too, though! One that's super exciting to me is the searchable broom method table—if you have a model object to summarize but can't find tidier methods for it, you can search across all CRAN packages here🤯
📣 Job with our clever, friendly, practical, well-funded team
NHS Senior Specialist Analyst - Data Engineering & Systems Development
£46-54k, 4 years, 95% remote
Join our team working to improve NHS care and reduce health inequalities. Post will focus on automating large-scale linkage of messy, high-security, cross-sector data
Happy to chat with anyone interested. We'll have a data science and a research fellow post open soon too
Hey #rstats hivemind... How does one go about either finding someone to take over maintenance of a (little-used) package or having that package archived? In the CRAN Policies, an author can request that a package by archived, but it is unclear how that request is made! If your interested the package is quickmapr and I have no plans to update it to sf/terra.
After some more fiddling I'm starting to get usable results with this Shiny app. I tried relying on the LLM to highlight edits it was suggesting, but it totally sucked at that. Now that I'm comparing differences using reliable functions (diffobj), it's easier to see what changes are being suggested. Still very cludgy, but I think I'll work on improving this a little bit. #rstats
If you maintain any #rspatial#rstats packages, you likely (hopefully!) know about the impending retirement of rgdal, rgeos, and rmaptools. This is a great writeup by @nowosad on how to transition to the modern sf/terra packages. Good for developers and users! https://geocompx.org/post/2023/rgdal-retirement/
I've been hinting about some larger things going on at work, and that I have not been in a position to say anything untill all applications were approved. It is not approved and I can say:
Me and two colleagues (as a start) are establishing a core facility at the University of Oslo to help researchers in social and humanities sciences with their projects.
And right now another message just popped in my inbox confirming more internal funding from my University for establishing a platform for online behavioural data collection, improving #rstats package installs for our airgapped server and creating a network to share competencies within social and humantity sciences with complex technological needs.
🥳 This is kind of the start of a new career within academia for me, and I am so excited!
Currently my R session will crash if I call usethis::use_version() and choose one of the option and hit enter. Has anyone in the #RStats@rstats experienced this?
I'm working on an internal charting package, based on the plots I created for our demographic/ adult health profiles.
This seems to involve a ton of boiler plate ggplot2 code and there must be someone, somewhere who has done something similar.
If so, please point me in that direction so I can think how best to structure the code
Note - this is not related to themes etc.
It's more the basics such a line chart template, a column chart template etc.
@johnmackintosh Probably not an actual answer to your question:
I have been down paths of creating custom plots with lots of arguments. It got painful quickly, as it diverged from what people already understood with {ggplot2}.
Instead I created a few custom geoms/stats, then it was just a case of telling people:
"You know ggplot. Here are our in-house custom geoms. Facets and everything else work the same. Go for it!"
I started at the end of 2012.
There's an online R cookbook, that was my Bible back then. That, and trying to decipher base R solutions on Stack Overflow.
I remember using {plyr} (not a typo, {dplyr} didn't exist) for the first time, to calculate medians, for lots of individual groups. Amazing.
ggplot2 - even back then, was streets ahead of anything else I'd used for graphics.
🫤 Unfortunately, I was affected by last week's layoffs at @Posit along with a handful of other talented colleagues.
🖤 I’m bummed, but so grateful for the 5½+ yrs w/ my incredible team.
🔍 Taking some time to recoup, but will be on the hunt for new opportunities soon!