Machine readable format

[rwohleb]

In addition to standardized field names ( #1 ), it would be great if all of the questions/answers were in an easily parsable format. I have some ideas for gathering demographics in a secure way, and would much prefer to be able to just slurp up everything already defined here, rather than having to manage that myself.

[rwohleb]

Since everything else is JSON these days, maybe it could be in a form such as:

age.json { "id": "age", "title": "Age", "description": "The age of the person", "options": [ {"id": "0to14", "title": "0 to 14"}, {"id": "15to19", "title": "15 to 19"}, { "id": "20to60", "title": "20 to 60", "options": [ {"id": "20to39", "title": "20 to 39"}, {"id": "40to60", "title": "40 to 60"} ] } ] }

This structure allows for different levels of granularity. One potential use is allowing people to broadly identify with a group without having to identify with a specific sub-group. When consuming the results, it would make it easy to roll up the stats without having to separately define a connection between values.

I'm exhausted from the end of a convention, so half of what I said may make no sense ;)

[drnikki]

Hi @rwohleb yes, this is great, thank you!! There have been some discussions of this offline - how can we enable form authors (who are also programmers) to just consume a json feed of the questions and then, theoretically, also stay up to date as questions evolve.

JSON is definitely what we'll end up using. As a first pass, can you suggest a way to structure this initiative in the issue queue?

[rwohleb]

I've started looking at what it would take to setup a PR for demonstration purposes. I'm currently trying to figure out how you are handling things between docs, doc-source, and gh-pages. Is there a README that I just happened to miss?

[drnikki]

@rwohleb nope - you caught me not documenting. :)

pip install sphinx
pip install sphinxcontrib-bibtex
pip install guzzle_sphinx_theme
sphinx-build -b html doc-source docs

it's sphinx-build that generates docs from doc-source

[drnikki]

I can add this to a readme if you can confirm that those steps work. :) that's all of the instances of sphinx that I could find in my bash history.

[hestenet]

Just wanted to chime in with a +1 for this issue. At DrupalCon Nashville I discussed this same idea briefly in the hallway with @drnikki and @sparklingrobots. Providing the questions and answers in JSON format would be incredibly helpful, and then my team over at Drupal.org can work on building a module to consume that for Drupal sites (and I can totally imagine other open source projects wanting to do the same).

Thanks for proposing this/moving it forward @rwohleb

[sparklingrobots]

Just a note to say that the DDI Contrib Team has been working on getting Gender Field module up for D8, which we are tying to OpenDemographics. Right now the module has a manually created version of the OpenDemographics data, but our goal is to build the module to consume this data.

We're quite close to getting this module out into beta and I think it would be swell if we could work together with you & the OpenDemographics folks to get this work done. I confess we already registered the namespace on d.o for OpenDemographics in anticipation of this work. :)

Posted this on the d.o issue queue as well.

[hestenet]

Thanks for the update @sparklingrobots! I'll get caught up on the Gender Field module work and start following the OpenDemographics project queue on D.O.

[ivanstegic]

I love the idea of there being a machine readable format.

I wonder if the full JSON response couldn't just be an endpoint of an HTTP request that is the latest released version on a static github pages URL, perhaps with a custom domain? That way, it would be easy to make that list available to anything that speaks HTTP (including mobile phone apps) while including versioning as well. Might be a good idea to have a separate repo for that? Perhaps even something that is owned by an open-demographics org? There's lots to unpack here... and maybe brainstorming it here is the wrong place, so I'm happy to do it elsewhere too.

[drnikki]

Hi all,

First, as always, thank you for participating in this! I think what makes the most sense as a next step is to agree on a data structure that would account for all questions and be extensible for the future. I've started a ticket for that here: https://github.com/drnikki/open-demographics/issues/38

[drnikki]

@rwohleb (can't tag you in #38 for some reason), but I did a pass on this and would love your feedback since you kicked it off (others too - @ivanstegic, @sparklingrobots <3 ) Thanks, all, for your continued help with this.