Since we've dropped EveryPolitician data in there, it would be lovely to explicitly link to everypolitician.org because we provide lots of potentially useful information both about the data and how to use it (e.g., that we have libraries for working with it/keeping it up to date)... which isn't obvious just from grabbing the data itself.

"More information:" maybe? Or (less hinty-helpy) "Website:"

Suggestion/example:

(oops no www needed on URL there)

Tasks to be completed before it is 'launched':

Add - Data & APIs Link to footer of:

• parser.theyworkforyou.com – @zarino / @wrightmartin – mysociety/parlparse#94
• research footer
• researchsites template (waiting on mailing list change)
• www.mysociety.org – @zarino / @wrightmartin

Add this text 'APIs and datasets for other mySociety services can be found on our data portal - http://data.mysociety.org/' to:

• parser.theyworkforyou.com – @zarino / @wrightmartin – mysociety/parlparse#94
• www.theyworkforyou.com/api – @zarino / @wrightmartin – mysociety/theyworkforyou#1350

Merge:

• Update to EveryPolitician docs - everypolitician/everypolitician-docs#41

Write:

• Blog Post on Data Portal
• Blog Post on EveryPolitician R Repo

Review (@MyfanwyNixon):

• Blog Post on Data Portal - https://blogs.mysociety.org/mysociety/wp-admin/post.php?post=41658&action=edit
• Blog Post on EveryPolitician R Repo - https://blogs.mysociety.org/mysociety/wp-admin/post.php?post=41659&action=edit

Publish:

• Blog Post on Data Portal
• Blog Post on EveryPolitician R Repo

Databot set a lot of timestamps an hour later then next day set them back: b2fe6ea
Dunno if this signifies a bug or not.

(spun out from #6 (comment))

The site currently requires JavaScript to show anything under /datasets/ and much of /sources/ even though the pages consist of basic HTML, which is being generated by a static site generator that could therefore be generating actual HTML for all these pages. Yet the site is not using JavaScript to do things I would expect it to e.g. filtering the lists page without a server-side request – each filter change causes a full page reload, which might as well then be loading the requested data.

The search could still work identically as an enhancement, that's exactly what JavaScript is for.

It makes the site slower (instead of just showing me the dataset or list, it shows an empty page that has to wait for the JS to download and fetch more data in order to show anything), it makes it less available - https://kryogenix.org/code/browser/why-availability/ - and it makes it less resilient - https://kryogenix.org/code/browser/everyonehasjs.html .

So that I'm not just a killjoy, I've pushed a branch that fixes all the individual /dataset/ pages and /source/ pages to actually have text on them, this was not much work (because in the first case it turns out the text was already there, it was just being hidden by default... 🤷‍♂️ , in the second I copied a JS template into the HTML and it Just Worked). Some more effort must be put in to do the same with:

• the various list pages (which could then have nicer URLs too)
• Only having the search box display if JS available

The footer on data.mysociety.org currently refers to UKCOD.

Fortunately, this appears to be quite trivial to rectify, as the data is contained in /_includes/mysoc-footer.html

This site currently uses a stock JKAN theme (it's Bootstrap with a Bootswatch slapped over the top), and it should probably look something like our other Jekyll-powered sites.