search

psych-ds / psych-DS

Welcome to Psych-DS! If this is your first time visiting a Github repository, look to the left/down to the README (below the repository files.) Psych-DS is a specification for behavioral datasets - JSON-LD metadata, predictable directory structure, and machine-readable specifications for tabular datasets in behavioral research
Creative Commons Attribution 4.0 International
77 stars 6 forks source link

SIPS 2019 hackathon plans #21

Closed mekline closed 4 years ago

mekline commented 5 years ago

Here is the abstract for the SIPS hackathon:

Less data cleaning, more data adventures with Psych-DS: What could we do if datasets lived in a common format?

Psych-DS (psych-ds.github.io) is a specification for datasets, designed to help researchers build consensus around machine-readable standards that work for both shared and protected data, and integrate with all our favorite tools. But a specification is only as good as its applications: at this hackathon, we'll have a brief (20-30 min) introduction to Psych-DS and some existing tools, and then dream up (or prototype!) the applications you want to see. Zero coding experience needed, just bring your ideas!

Some things to figure out:

mekline commented 5 years ago

Here's a google doc to organize some planning for this session!

https://docs.google.com/document/d/1MpjY5PU4M3YNp5UoXnEKzif2AWp3DUg5qzK6Xcg-yew/edit?usp=sharing

mekline commented 5 years ago

One thing I've noticed at previous events like this is the 'problem' of people suggesting tools that already exist, or that are similar to ongoing efforts. This should be great news, because it means there is a lot of interest, but it can feel kind of deflating (or even, you can tell someone that their thing already exists, when actually they have a new/different idea). I'd love to talk about how we can balance discovering what people dream up on their own with helping people find existing tools. (Of course, "An X for Psych-DS" for almost any value of X is going to be a needed thing, at the moment!)

debruine commented 5 years ago

This might be tackled by focussing on things we want to do with datasets and parking the issue of whether Psych-DS should do the thing itself or use an existing solution for any actual implementation. If I understand corectly, this hackathon is aimed at brainstorming applications more than implementing them, right?

schnarrd commented 5 years ago

I think we should dedicate a fair amount of time not just to figuring out good tools to build on top of Psych-DS, but also figuring out how to lower barriers to adoption. Psych-DS involves some unfamiliar file types (.tsv and .json), and, in my opinion, we'll need some good tech writing that describes what the components of Psych-DS are (in the plainest possible language) and why someone would want to use Psych-DS at all. Some of this is there, but in order to get good uptake on the standard, these will have to be very clear and persuasive, I think.

On Fri, Apr 5, 2019 at 6:54 AM Lisa DeBruine notifications@github.com<mailto:notifications@github.com> wrote:

This might be tackled by focussing on things we want to do with datasets and parking the issue of whether Psych-DS should do the thing itself or use an existing solution for any actual implementation. If I understand corectly, this hackathon is aimed at brainstorming applications more than implementing them, right?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/psych-ds/psych-DS/issues/21#issuecomment-480248308, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AFS5gTwe6jTP-hfXbNXnkDT3dBrolFMoks5vdzlegaJpZM4cd8FA.

-- Patrick S Forscher Assistant Professor Department of Psychological Science University of Arkansas

Website: forscher.uark.eduhttp://forscher.uark.edu

chartgerink commented 5 years ago

Heyo! :wave:

I am joining at SIPS and love the idea of creating a data spec.

I've been inspecting Frictionless data from the Open Knowledge foundation, and stumbled on this on Twitter. I'd be happy to contribute and have some programming experience. Maybe also see how to make these specs cross-compatible where possible, or writing export tools.

Well, just dropping a line to hope and get involved :)