joshdeleeuw.bsky.social
Cognitive Scientist at Vassar College. Currently working on open source tools for researchers.
Creator and maintainer: https://www.jspsych.org
57 posts
344 followers
330 following
Regular Contributor
Active Commenter
comment in response to
post
I teach a methods class annually where this could be a huge help.
comment in response to
post
👋 I'm doing lots of RSE work these days
comment in response to
post
You could have scenarios where the value and keys are different vectors (incl. different sizes) coming from different sources. This isn't the common use case, but the general mechanism allows it.
comment in response to
post
fwiw I found the Rmarkdown -> Quarto transition to be smooth. There's a lot of overlap.
comment in response to
post
whoa this is really interesting - my dissertation work was on similar kinds of mechanisms and models but in the statistical learning world. I never published it, but it's here osf.io/5wdnb
congrats on finishing!
comment in response to
post
Please share with folks who might be interested in participating!
comment in response to
post
If none of the available times work, you can fill out this form and we'll reach out to schedule a time.
docs.google.com/forms/d/e/1F...
comment in response to
post
If you are interested, you can reserve a time on this calendar.
calendar.google.com/calendar/u/0...
comment in response to
post
Can we implement and add to our new shareable timelines project for jsPsych?
github.com/jspsych/jsps...
comment in response to
post
hey @willngiam.bsky.social we are working on something pretty similar to this in my group right now and would be happy to collaborate. feel free to DM/email me if you want to chat more in depth!
comment in response to
post
thanks. this is how I do it to but my knife skills aren't good enough to call this "quick" 😅
comment in response to
post
sam now that you are an expert what is the fast way to peel kiwis?
comment in response to
post
This is a cool case study in this sort of thing: philarchive.org/rec/MUREAE-6
comment in response to
post
It's easy...ish.
You can create factorial combinations with www.jspsych.org/latest/refer...
Then pass that list in as a set of timeline variables:
www.jspsych.org/latest/overv...
And finally use a sampling process to pick a subset:
www.jspsych.org/latest/overv...
comment in response to
post
the dude in the middle seat must have been real uncomfortable
comment in response to
post
This release also sets us up to fully support the Psych-DS standard. We're working on a set of tools that will automatically generate metadata and move us closer to machine-readable datasets.
comment in response to
post
Here's the migration guide from v7 to v8. This is a much simpler transition than v6 to v7. Most v7 experiments will probably work right away with v8.
www.jspsych.org/v8/support/m...
comment in response to
post
DataPipe is a free tool for saving experimental data directly to the OSF, making it born-open by default. pipe.jspsych.org doi.org/10.31234/osf... @joshdeleeuw.bsky.social
comment in response to
post
niche, but 👍
comment in response to
post
Yeah, totally in agreement! "Possible but clunky" is not a great developer experience 😅
comment in response to
post
And along the lines of "consistent API" I think a related advantage is something analogous to an advantage of the tidyverse in R. If you have a consistent way to construct an analysis pipeline then you can build tools that interact easily, without needing centralized development.
comment in response to
post
I think the restrictiveness that advanced users sometimes feel is more of a developer experience fault on my end than an actual restrictiveness. I've yet to encounter something that couldn't be implemented. (Which is not to say that it is a good idea to use it for every project!)
comment in response to
post
Not quite as beautiful, but a vscode extension can transform the editor: e.g. github.com/zaaack/vscod...
comment in response to
post
comment in response to
post
and if you want super cheap combine it with pipe.jspsych.org and you can run the entire thing for free
comment in response to
post
I saw my first yesterday. would totally schedule a trip to include this.
comment in response to
post
what a coincidence I am available for jspsych workshops in 2028
comment in response to
post
I have a student working on memory enhancement in VR for her bachelor's thesis. Are non-UK students eligible?
comment in response to
post
Hmmm good point. I don't know either! The cloud function doesn't store any long-term record of the payload, but I don't know if that's sufficient.
comment in response to
post
No built-in method for this yet, but it's worth adding.
I think this will work if you stick it in an `on_finish`-type spot:
document.querySelector("#jspsych-progressbar-container span").innerHTML = "new message";
comment in response to
post
well now I need to know what other anti-design patterns we can add to the library...