Psych-DS example datasets & (proto) validator!

[melissa...@gmail.com]

Dear all - it's been quite a while since I've had any reason to write to this group (and my apologies for those who see very similar content in a twitter thread :)  )

These developments principally come in the form of two things that popped up in just the last few months (and which I've been falling into the trap of "getting ready" with before asking for help and input!)

First, I scheduled a hackathon at SIPS, planning to work with attendees to build some example Psych-DS datasets by hand - we've tried this before, and it's been my sense that not having many concrete examples is a barrier both to helping people understand Psych-DS, and to tool-builders who might like to work with PDS data.

Literally on the plane, I decided to see how far I could get with a shinyApp validator that does some pretty minimal validation of file name and folder structure, and then sends the user over to https://validator.schema.org/ to paste in the contents of dataset_description.json. Here is the result. 

We've discussed previously whether or not Psych-DS should adopt the full JSON-LD Dataset model before - for the purposes of this hackathon, I went with it specifically because a tool existed for checking. There's been concern about users being able to work with that file, but we had some great successes with JSON novices using the Schema validator to make minor changes to a minimal dataset_description.json example.

Similarly, this validator tool implements CSV rather than TSV data encoding because it made it easier for hackathon participants to start converting an existing dataset.

The second recent development was an opportunity to write a grant for the NIH Brain initiative, in which I proposed funding to write *real* validator software, implement it in some existing open source experiment-presentation tools (jsPsych, psychoPy, Lookit), and define the BIDS/Psych-DS translation.  I wrote it very much at the last minute, but it's what got me motivated to pick this project up again after a few years at a job where I couldn't work on it :) 

I know this is a bunch of information, and I think the correct next step would be to schedule a meeting to chat about progress, go over existing tools & efforts, and identify some goals to work toward. I've created a placeholder agenda doc for people to begin adding items to, and will follow up early next week to start the scheduling process. 

Thank you all for your hard work and support on this project already - if you'd be interested in getting involved in any type of way during the next several months, please be in touch!!

All the best, 

Melissa