Aslan B.
@motorneurosci.bsky.social
PhD Student @ Donders Institute. Interests: a bit of information processing in the human sensorimotor system, a big share of open reproducible science, a nonsignificant amount of stats, a limited amount of causal inference, and a heavy dose of Olympic WL.
Projects in {targets} (books.ropensci.org/targets/proj...) is a game changer for running part of the pipeline on HPC. Declare the targets from the non-HPC store as dependencies in the HPC pipeline once and use tar_cue_skip. Only need to copy / git track the HPC target store, not everything. #rstats
8 Projects – The {targets} R package user manual
books.ropensci.org
October 16, 2025 at 3:37 PM
Projects in {targets} (books.ropensci.org/targets/proj...) is a game changer for running part of the pipeline on HPC. Declare the targets from the non-HPC store as dependencies in the HPC pipeline once and use tar_cue_skip. Only need to copy / git track the HPC target store, not everything. #rstats
Reposted by Aslan B.
Today I'm celebrating the 200th github ⭐ for legendry after talking about it yesterday to the ggplot2 extension group!
teunbrand.github.io/legendry/
teunbrand.github.io/legendry/
Extended Legends and Axes for ggplot2
A ggplot2 extension that focusses on expanding the plotters arsenal of guides. Guides in ggplot2 include axes and legends. legendry' offers new axes and annotation options, as well as new legends and ...
teunbrand.github.io
January 22, 2025 at 5:47 PM
Reposted by Aslan B.
Here is something I suspect happens *a lot*:
Researchers have nested data.
Run rmANOVA.
Don't get significant results.
"Hey, how about multilevel models?"
Run MLM with only random intercept.
Everything is significant.
"I guess MLMs have more power!"
Publish.
#stats
Researchers have nested data.
Run rmANOVA.
Don't get significant results.
"Hey, how about multilevel models?"
Run MLM with only random intercept.
Everything is significant.
"I guess MLMs have more power!"
Publish.
#stats
December 23, 2024 at 10:31 AM
Here is something I suspect happens *a lot*:
Researchers have nested data.
Run rmANOVA.
Don't get significant results.
"Hey, how about multilevel models?"
Run MLM with only random intercept.
Everything is significant.
"I guess MLMs have more power!"
Publish.
#stats
Researchers have nested data.
Run rmANOVA.
Don't get significant results.
"Hey, how about multilevel models?"
Run MLM with only random intercept.
Everything is significant.
"I guess MLMs have more power!"
Publish.
#stats
Reposted by Aslan B.
New preprint with @jdcrawford.bsky.social
We suggest that the old serial transformation story in motor planning needs updating: intrinsic (muscle like) motor plans seem to appear BEFORE extrinsic (spatial) plans... 🧠
www.biorxiv.org/content/10.1...
We suggest that the old serial transformation story in motor planning needs updating: intrinsic (muscle like) motor plans seem to appear BEFORE extrinsic (spatial) plans... 🧠
www.biorxiv.org/content/10.1...
MEG signals reveal arm posture coding and intrinsic movement plans in parietofrontal cortex
Movement planning processes must account for body posture to accurately convert sensory signals into movement plans. While movement plans can be computed relative to the world (extrinsic), intrinsic m...
www.biorxiv.org
December 3, 2024 at 1:58 AM
New preprint with @jdcrawford.bsky.social
We suggest that the old serial transformation story in motor planning needs updating: intrinsic (muscle like) motor plans seem to appear BEFORE extrinsic (spatial) plans... 🧠
www.biorxiv.org/content/10.1...
We suggest that the old serial transformation story in motor planning needs updating: intrinsic (muscle like) motor plans seem to appear BEFORE extrinsic (spatial) plans... 🧠
www.biorxiv.org/content/10.1...
Reposted by Aslan B.
I can't believe John Ioannidis is still a thing! 😂
The reproducibility crisis and other problems in science – John Ioannidis
https://www.youtube.com/watch?v=vY9mGJQFdyE
https://www.youtube.com/watch?v=vY9mGJQFdyE
December 1, 2024 at 1:44 AM
I can't believe John Ioannidis is still a thing! 😂
Reposted by Aslan B.
I now dedicate the first 20+ lines of my user facing functions to argument validation.
I cannot emphasize how helpful this is for preventing bugs.
Read about r-lib standalone check: josiahparry.com/posts/2024-0...
#rstats
I cannot emphasize how helpful this is for preventing bugs.
Read about r-lib standalone check: josiahparry.com/posts/2024-0...
#rstats
November 26, 2024 at 7:34 PM
I now dedicate the first 20+ lines of my user facing functions to argument validation.
I cannot emphasize how helpful this is for preventing bugs.
Read about r-lib standalone check: josiahparry.com/posts/2024-0...
#rstats
I cannot emphasize how helpful this is for preventing bugs.
Read about r-lib standalone check: josiahparry.com/posts/2024-0...
#rstats
Reposted by Aslan B.
𝚝𝚒𝚗𝚢𝚝𝚊𝚋𝚕𝚎 0.6.1 for #Rstats is out!
It's an ultra simple, super flexible, and 0-dependency package to draw beautiful tables in HTML, LaTeX, Typst, Word, PDF, and PNG.
And for those who ❤️ documentation, 𝚝𝚒𝚗𝚢𝚝𝚊𝚋𝚕𝚎 ships with a billion pages of tutorials:
vincentarelbundock.github.io/tinytable/
It's an ultra simple, super flexible, and 0-dependency package to draw beautiful tables in HTML, LaTeX, Typst, Word, PDF, and PNG.
And for those who ❤️ documentation, 𝚝𝚒𝚗𝚢𝚝𝚊𝚋𝚕𝚎 ships with a billion pages of tutorials:
vincentarelbundock.github.io/tinytable/
November 21, 2024 at 2:38 PM
𝚝𝚒𝚗𝚢𝚝𝚊𝚋𝚕𝚎 0.6.1 for #Rstats is out!
It's an ultra simple, super flexible, and 0-dependency package to draw beautiful tables in HTML, LaTeX, Typst, Word, PDF, and PNG.
And for those who ❤️ documentation, 𝚝𝚒𝚗𝚢𝚝𝚊𝚋𝚕𝚎 ships with a billion pages of tutorials:
vincentarelbundock.github.io/tinytable/
It's an ultra simple, super flexible, and 0-dependency package to draw beautiful tables in HTML, LaTeX, Typst, Word, PDF, and PNG.
And for those who ❤️ documentation, 𝚝𝚒𝚗𝚢𝚝𝚊𝚋𝚕𝚎 ships with a billion pages of tutorials:
vincentarelbundock.github.io/tinytable/
Pr( written in rust | described as "blazingly fast") = 1
November 18, 2024 at 11:39 AM
Pr( written in rust | described as "blazingly fast") = 1
Reposted by Aslan B.
Good discussion on brain decoding, modularity and lots more.
#neuroscience
#neuroscience
November 16, 2024 at 8:49 PM
Good discussion on brain decoding, modularity and lots more.
#neuroscience
#neuroscience
Reposted by Aslan B.
1/ X's algorithm was changed in mid-July 2024 to systematically boost Republican-leaning accounts and Elon Musk's own account following his endorsement of Donald Trump, according to a newly released computational study of engagement from the Queensland University of Technology.⬇️
November 16, 2024 at 10:40 AM
1/ X's algorithm was changed in mid-July 2024 to systematically boost Republican-leaning accounts and Elon Musk's own account following his endorsement of Donald Trump, according to a newly released computational study of engagement from the Queensland University of Technology.⬇️
New TIL for {targets} #rstats: For static branching, you'd use tarchetypes::tar_map. But what if you want to assign different resources to those branches? You cannot do this with tar_map (please correct if wrong). Fortunately, you can create you list of targets in any way, including purrr::map... >
November 15, 2024 at 4:17 PM
New TIL for {targets} #rstats: For static branching, you'd use tarchetypes::tar_map. But what if you want to assign different resources to those branches? You cannot do this with tar_map (please correct if wrong). Fortunately, you can create you list of targets in any way, including purrr::map... >
Still on my radar. One day...
Is there {targets} + {crew} for Julia? Looks like Pipelines.jl + JobScheduler.jl might be it?
Is there {targets} + {crew} for Julia? Looks like Pipelines.jl + JobScheduler.jl might be it?
Hi #rstats Twitter. Since many of you were on other platforms for a while, wanted to share that we have been busy building Tidier.jl for #JuliaLang.
The goal? All of tidyverse, written in 100% Julia.
Check it out: tidierorg.github.io/Tidier.jl/dev/
JuliaCon 2023 intro: youtu.be/ZVvP7rAIvkE?...
The goal? All of tidyverse, written in 100% Julia.
Check it out: tidierorg.github.io/Tidier.jl/dev/
JuliaCon 2023 intro: youtu.be/ZVvP7rAIvkE?...
Tidier.jl
A VitePress Site
tidierorg.github.io
November 15, 2024 at 6:38 AM
Still on my radar. One day...
Is there {targets} + {crew} for Julia? Looks like Pipelines.jl + JobScheduler.jl might be it?
Is there {targets} + {crew} for Julia? Looks like Pipelines.jl + JobScheduler.jl might be it?
If this ends up in base R it would definitely make the language quite a bit nicer imo (syntax wise but of course also nice to have just one system for each functional and encapsulated oop).
S7 0.2.0 is now on CRAN, thanks primarily to the hard work of @t-kalinowski.bsky.social. S7 is a collaborative effort to supersede S3 and S4 and you can read all about it (and what's new in 0.2.0) at www.tidyverse.org/blog/2024/11... #rstats
S7 0.2.0 - Tidyverse
S7 is a new package that simplifies object-oriented programming (OOP) in R. It combines the simplicity of S3 with the structure of S4 to create a clearer system that's accessible to everyone.
www.tidyverse.org
November 12, 2024 at 9:04 PM
If this ends up in base R it would definitely make the language quite a bit nicer imo (syntax wise but of course also nice to have just one system for each functional and encapsulated oop).
#rstats {targets} tip: when rewriting dynamic branches as static branches via tar_map, use tibble::lst( mapped_targets = tar_map(...), tar_combine(combined_name, tar_select_targets(mapped_targets, ))) to easily combine the branches.
November 10, 2024 at 10:51 PM
#rstats {targets} tip: when rewriting dynamic branches as static branches via tar_map, use tibble::lst( mapped_targets = tar_map(...), tar_combine(combined_name, tar_select_targets(mapped_targets, ))) to easily combine the branches.
Answer now here: github.com/ropensci/tar...
For me, this is a good argument to prefer static over dynamic branching as a default in {targets} #rstats
For me, this is a good argument to prefer static over dynamic branching as a default in {targets} #rstats
November 10, 2024 at 3:24 PM
Answer now here: github.com/ropensci/tar...
For me, this is a good argument to prefer static over dynamic branching as a default in {targets} #rstats
For me, this is a good argument to prefer static over dynamic branching as a default in {targets} #rstats
Can someone explain this {targets} behaviour: with dynamic branching, targets whose ancestors have not changed rerun? #rstats
stackoverflow.com/questions/79...
stackoverflow.com/questions/79...
targets dispatch vs. skip behaviour is altered with dynamic branching
The code below illustrates it. There are three targets dfr -> dfr_x -> result. If i change dfr$a, it does not affect the downstream dfr_x and hence not result. Yet, with dynamic branching, all
stackoverflow.com
November 10, 2024 at 9:37 AM
Can someone explain this {targets} behaviour: with dynamic branching, targets whose ancestors have not changed rerun? #rstats
stackoverflow.com/questions/79...
stackoverflow.com/questions/79...
#rstats {targets} & quarto parametrized reports: say I have a list of plots, and a parametrized report that shows one plot based on a index parameter. What happens is that each report has the same plot, because the plot file will get overwritten >
November 6, 2024 at 10:28 PM
#rstats {targets} & quarto parametrized reports: say I have a list of plots, and a parametrized report that shows one plot based on a index parameter. What happens is that each report has the same plot, because the plot file will get overwritten >
Reposted by Aslan B.
Bluesky is great, but the only thing I want on my social media other than research threads is academic shitpoasting.
November 5, 2024 at 11:12 PM
Bluesky is great, but the only thing I want on my social media other than research threads is academic shitpoasting.
Reposted by Aslan B.
figuring out {targets} #rstats on Northwestern's HPC, part 617:
if your targets are segfaulting randomly, they're probably being killed for using too much RAM since targets passes all data through the login node by default, change it to have workers read/write data directly
if your targets are segfaulting randomly, they're probably being killed for using too much RAM since targets passes all data through the login node by default, change it to have workers read/write data directly
November 3, 2024 at 8:54 PM
figuring out {targets} #rstats on Northwestern's HPC, part 617:
if your targets are segfaulting randomly, they're probably being killed for using too much RAM since targets passes all data through the login node by default, change it to have workers read/write data directly
if your targets are segfaulting randomly, they're probably being killed for using too much RAM since targets passes all data through the login node by default, change it to have workers read/write data directly
woah woah woah you can bring declarative NixOS style management to your Macbook??? Awesome
github.com/LnL7/nix-dar...
github.com/LnL7/nix-dar...
GitHub - LnL7/nix-darwin: nix modules for darwin
nix modules for darwin. Contribute to LnL7/nix-darwin development by creating an account on GitHub.
github.com
October 31, 2024 at 12:26 PM
woah woah woah you can bring declarative NixOS style management to your Macbook??? Awesome
github.com/LnL7/nix-dar...
github.com/LnL7/nix-dar...