Jonathan Carroll
LightNews LightNews
banner
jonocarroll.fosstodon.org.ap.brid.gy
Jonathan Carroll
@jonocarroll.fosstodon.org.ap.brid.gy
98 followers 3 following 75 posts
#rstats primarily but learning #haskell, #julia, #rust, #python, #APL, #lisp, and many others

:rstats: :haskell: :julia: :rust: :apl: :python: :bash:

Also […]

🌉 bridged from ⁂ https://fosstodon.org/@jonocarroll, follow @ap.brid.gy to interact
Posts Replies Media Videos
jonocarroll.fosstodon.org.ap.brid.gy
About half way through @daroczig's #user talk recording on {logger}

https://www.youtube.com/watch?v=8R8ulz4MVm0

I was admiring how extensible it was and it struck me that it was a good fit for sending logged errors via {ntfy} - I got a working proof of […]

[Original post on fosstodon.org]
logger message sent via ntfy
October 31, 2025 at 4:14 AM
jonocarroll.fosstodon.org.ap.brid.gy
The {ntfy} #rstats package has been updated on CRAN with internal improvements and tests thanks to @hadleywickham https://cran.r-project.org/package=ntfy

Send a message to your devices on-demand or when some code completes, including emoji and plot outputs. No […]

[Original post on fosstodon.org]
ggplot sent as a notification via the ntfy package
October 26, 2025 at 9:30 PM
jonocarroll.fosstodon.org.ap.brid.gy
Got a good bird-to-stone ratio going by refreshing my #neo4j #cypher knowledge and warming up for #adventofcode by solving 2024 day 10 entirely via graph queries (including parsing input). I think my theme for this year is going to be 'as many languages as possible'
Neo4j Desktop showing part of an advent-of-code solution
October 26, 2025 at 2:00 AM
jonocarroll.fosstodon.org.ap.brid.gy
@zachweinersmith.bsky.social Thank You For Smoking is my go-to example of "winning the argument for the wrong reason" https://www.moviequotedb.com/movies/thank-you-for-smoking/quote_23377.html

My guess is that a 2hr debate is not a format in which you need to […]

[Original post on fosstodon.org]
And you should pay attention to your confusion, because either your belief model is wrong, or something else is going on that you're not aware of. It is a design flaw in human cognition that this sensation manifests as a quiet strain in the back of your mind, instead of a wailing alarm siren and a glowing neon sign reading: Either Your Model Is False Or This Story Is Wrong! If you're trying to become a more rational being - a better Rationalist - then you need to listen to those little moments when something doesn't quite seem to add up.
October 24, 2025 at 11:55 PM
jonocarroll.fosstodon.org.ap.brid.gy
I Vibe Coded an R Package
and it ... actually works?!?

https://jcarroll.com.au/2025/09/13/i-vibe-coded-an-r-package/

I wanted an #rstats package to exist so I spent $20 and got Claude Code to build it for me.

Plus a shiny app using it: https://jcarroll.shinyapps.io/kanjiexplorer/
Kanji Explorer available at https://jcarroll.shinyapps.io/kanjiexplorer/
September 13, 2025 at 1:41 AM
jonocarroll.fosstodon.org.ap.brid.gy
I'm not sure where the #user2025 discussions are happening but I'm curious about one slide in particular from @hadleywickham's talk

https://www.youtube.com/live/ctc2kx3LxG8

why the addition of 0 for each term of the multiplication? Is it to have NA values […]

[Original post on fosstodon.org]
Slide from Hadley's talk showing (x+0)*(y+0)
August 10, 2025 at 12:08 PM
jonocarroll.fosstodon.org.ap.brid.gy
What are the odds? @dlakelan @jimbob

Would only have been better with the latter's previous username, or if the supplied landscape picture was of a lake, someone else named Daniel, or Daniel at a lake.

😳
A pun on "Bela Lugosi's Dead" followed by a post from account with a different pun of the same as their username (who previously had exactly that first pun as their username, IIRC)
July 21, 2025 at 8:44 AM
jonocarroll.fosstodon.org.ap.brid.gy
Fractional exponents are a thing

x ^ z

where z is not an integer. For example x ^ 0.5 is the same as taking the square root, sqrt(x).

I must have either not learned, or forgotten that this only applies to positive values of x and spent a bit of time […]

[Original post on fosstodon.org]
Owlturd 'Yeah That Makes Sense' meme showing 2^3 = 8 (-2)^3 = -8 2^3.1 = 8.574 (-2)^3.1 = NaN
June 19, 2025 at 7:44 AM
jonocarroll.fosstodon.org.ap.brid.gy
A brilliant talk from Aaron Hsu (not on Fedi, AFAICT) https://youtu.be/V8sACAhg4vM

I'm left wondering though, is his lack of satisfaction in types due to #apl having so few of them? I absolutely see the benefits of typing things more strictly, and at the same […]

[Original post on fosstodon.org]
June 15, 2025 at 9:28 AM
jonocarroll.fosstodon.org.ap.brid.gy
{ggeasy} has now been updated to 0.1.6 on CRAN https://cran.r-project.org/package=ggeasy to align to an upcoming {ggplot2} release, but in fixing the revdep issue I discovered that easy_labs() (which used column labels as axis labels if present as attributes) […]

[Original post on fosstodon.org]
June 15, 2025 at 6:49 AM
jonocarroll.fosstodon.org.ap.brid.gy
Help me, #rstats fans: bug, or not a bug?

It looks like format(x, "%s") ignores the set timezone on a Date and assumes it's local, converting back to UTC even if it is explicitly already UTC. Should it ... not do that?

I got bitten by this because it didn't […]

[Original post on fosstodon.org]
# date locally local <- as.POSIXct("2025-06-04") # date in UTC utc <- as.POSIXct("2025-06-04", tz = "UTC") # convert to integer (a <- as.integer(utc)) #> [1] 1748995200 # number of seconds since the epoch (b <- as.integer(format(utc, "%s"))) #> [1] 1748961000 # difference (a - b) / 60 / 60 #> [1] 9.5 utc - local #> Time difference of 9.5 hours
June 4, 2025 at 7:59 AM
jonocarroll.fosstodon.org.ap.brid.gy
I'm at the R IDE. I'm at the Python IDE. I'm at the combination R and Python IDE.
Positron IDE welcome logo
May 10, 2025 at 1:17 PM
jonocarroll.fosstodon.org.ap.brid.gy
It might survive long enough to have a 10th birthday next year, but my silly little #rstats :rstats: package {ggghost} has been updated on CRAN to fix some broken documentation (and long-overdue cleanup)

https://cran.r-project.org/package=ggghost

I really […]

[Original post on fosstodon.org]
April 15, 2025 at 7:24 AM
jonocarroll.fosstodon.org.ap.brid.gy
Part book review, part package review, part comparison between :rstats: and :python: in my latest blog post

https://jcarroll.com.au/2025/02/13/book-review-pandas-workout/

Adelaide - we know all about "import pandas" 🐼
"import pandas" A newly arrived panda at the Adelaide Zoo, on loan as part of a breeding program.
February 12, 2025 at 11:44 PM
jonocarroll.fosstodon.org.ap.brid.gy
I'll be at EverythingOpen2025 this week #eo2025 #everythingopen looking to chat with anyone interested in data, programming languages (esp. R, Rust, Haskell, Julia, APL), or just about anything you're in to.

I plan to have my profile pic on my badge - say hi if you see me!
My profile picture, a cartoon rendering of me
January 19, 2025 at 1:50 AM
jonocarroll.fosstodon.org.ap.brid.gy
Interesting to try out the DataFrame module in #haskell :haskell: with some #dplyr familiarity https://github.com/mchav/dataframe - it's definitely not somewhere I'd start for analysing data, but I've been curious as to whether or not it was even possible with strict typing. It looks like it is!
Demonstrating usage of the Haskell Data.DataFrame module exploring some aggregations of a starwars dataset
January 9, 2025 at 4:58 AM
jonocarroll.fosstodon.org.ap.brid.gy
@bthalpin I found this example to be very interesting and wanted to see if I could golf it down a little with Distances.jl - I think it come out quite clean.

https://gist.github.com/jonocarroll/8180c068fa46fb3a3be4d7bbf0fd146f
January 9, 2025 at 2:13 AM
jonocarroll.fosstodon.org.ap.brid.gy
I love #rstats :rstats:, and believe that vectorised operations simplify code by making x + y do what it looks like it should.

#julialang :julia: has a more explicit approach; a 'broadcasting' operator `.` to distinguish scalar and vector operations, so x .* […]

[Original post on fosstodon.org]
ghci> let x = [1, 2, 3] ghci> let y = [4, 5, 6] ghci> x + y <interactive>:3:1: error: [GHC-39999] • No instance for ‘Num [Integer]’ arising from a use of ‘it’ • In the first argument of ‘print’, namely ‘it’ In a stmt of an interactive GHCi command: print it ghci> import Linear ghci> x ^+^ y [5,7,9]
December 16, 2024 at 10:17 AM
jonocarroll.fosstodon.org.ap.brid.gy
Just once, could I please read a book by an American who doesn't assume that I, the reader, is American?

There are - I'm am told - more than a few of us.
Extract from The Loop: "In fact, just standing in court, your or I would likely be under the gaze of an algorithm. [...] are in use in forty-six states and DC [...]"
December 14, 2024 at 12:23 PM