datasets / publicbodies

A database of public bodies such as government departments, ministries etc.
http://publicbodies.org
MIT License
63 stars 26 forks source link

Adding publicbodies as IATI codelist #74

Closed danfowler closed 2 years ago

danfowler commented 8 years ago

As suggested by @timgdavies here, we should submit a code for PublicBodies.org to IATI's Organisation Registration Agency codelist.

Related: https://github.com/okfn/publicbodies/issues/41

Something like this?

Code Name Description Category URL Public Database?
XI-PB Public Bodies Public Bodies aims to provide a unique ID for every part of every government. XI http://publicbodies.org/ Yes

How to do it:

If a Registation Agency that you wish to use is not on this list, please contact support@iatistandard.org

danfowler commented 8 years ago

@rgrp thoughts?

timgdavies commented 8 years ago

+1

bill-anderson commented 8 years ago

Will forward this to IATI Tech Team. Takes 2 weeks to process (I think)

rorads commented 8 years ago

Hello all,

This looks great! I've just added it to our non-embedded codelist consultations - see here.

I'm just thinking - when constructing an IATI Ogranisation Identifier, the regex [^\/\&\|\?]+ has to be satisfied. This was partly done so that IATI identifiers would play nice with URIs/URLs.

One simple fix would be for publishers to use a dash or underscore instead of a slash to concatenate the country code to the id.

So where an entry on Public Bodies reads:

id name abbreviation other_names description ...
gb/2010-rotherham-limited 2010 Rotherham Limited ... ... ... ...

The IATI Identifier for the corresponding organisation could be XI-PB-GB-2010-rotherham-limited or XI-PB-GB_2010-rotherham-limited.

The only issue with this is that will hinder lookup/searching from Public Bodies to IATI or vice versa, and secondary reporters would have to find out or decide how to convert between the two. Does anyone have any proposals of how best to deal with this?

I imagine it would be quite a big ask for you the replace the slash with an underscore, or something similar, but for anyone wanting to use an ID from this database in a URL it would certainly be helpful.

Just a thought. I look forward to seeing this develop - I expect it to be very helpful for data users in IATI or otherwise.

Thanks!

rufuspollock commented 8 years ago

@rory09 how does IATI (or specs that IATI references) generally handle "country code + in country identifier" stuff? The reason for the gb/ was actually so that it worked in urls in that you had the country then the within country identifier?

rorads commented 8 years ago

Hi @rgrp - that's interesting. The central difference with IATI is that we use identifiers as string references within XML attributes. So, for instance, if a publisher were to declare a participating organisation within an IATI activity, they would want to include the following line of XML:

  <participating-org ref="GB-COH-123456789" role="1" type="40"> 
  <!-- the role and type values are declared in codelists online, along 
       with the codes for country and registration agency -->
    <narrative>Name of Agency</narrative>
  </participating-org>

here, the ref="..." is where the identifier is declared. The above follows the structure "<country code>-<agency code>-<agency assigned identifier>", in this case Great Britain, Companies House, and digits 1-9.

So the short answer is that IATI handles this kind of concatenation with dashes. The motivation for this was, I think, partly to do with the variety of identifiers that would be dealt with (not all would have a country code), as well as wanting the identifier to be easily inserted into a url as is, given the variety.

I hope this is useful context - if you'd like to know any more about IATI, please let me know. For now, I've scheduled for Public Bodies to be included towards the end of the next week. If you have any further ideas for how best to construct IATI identifiers given the above, please let me know.

todrobbins commented 7 years ago

@rory09 where did we leave off on this? Was Public Bodies included last year?

Cheers!

timgdavies commented 7 years ago

Public bodies is in the http://org-id.guide codelist that will shortly replace the IATI list (http://org-id.guide/list/XI-PB)

I don't think the '/' character substitution issue has been resolved yet.

todrobbins commented 7 years ago

Thanks for the update @timgdavies.

VladimirAlexiev commented 7 years ago

@timgdavies "http://org-id.guide codelist will shortly replace the IATI list": are you speaking on behalf of IATI?

@todrobbins XI-PB is included in http://iatistandard.org/202/codelists/OrganisationRegistrationAgency/, so you can close this issue

The "slash vs underscore" issue reflects a big difference in philosophy, so I believe it merits a new issue to be opened: https://github.com/IATI/IATI-Codelists-NonEmbedded/issues/152

"Does this mean when an official register becomes available, we should deprecate "aggregator" identifier schemes or URLs (like OpenCorporates or PublicBodies) in favor of that official register?" (please comment there)

timgdavies commented 7 years ago

@VladimirAlexiev I don't speak for IATI - but the issue where this change is proposed and so far has had no objections can be found at https://discuss.iatistandard.org/t/migration-of-organisationregistrationagency-codelist-to-org-id-guide/851

VladimirAlexiev commented 7 years ago

Thanks for the info!

rorads commented 7 years ago

Apologies all, this is my personal email and I think these emails have been falling foul of an overzealous filter I set up to deal with github notifications.

Thank you Tim for clearing this up. I should mention that whilst I used to be a part of the IATI secretariat, I'm no a member of Open Data Services Co-operative along with Tim, so I also don't speak for IATI. That being said, I've worked with Tim in collaboration with the IATI Secretariat on the org-id.guide project, and I would agree that it is very likely to become the official organisation-id prefix list.

Thanks for setting up the new ticket, Vladimir - I'll see you in the comments section there!

Best, Rory

On 22 May 2017 at 10:08, Vladimir Alexiev notifications@github.com wrote:

Thanks for the info!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/okfn/publicbodies/issues/74#issuecomment-303043401, or mute the thread https://github.com/notifications/unsubscribe-auth/AHcY8dEQzkBtAyNS-UXDTQYlx3GM14Ygks5r8VCSgaJpZM4HOO0M .

augusto-herrmann commented 6 years ago

@todrobbins XI-PB is included in http://iatistandard.org/202/codelists/OrganisationRegistrationAgency/, so you can close this issue

Indeed it included there. @VladimirAlexiev thank you for pointing it out.

However, the line has some information that I think is incorrect. In particular the Category and Public Database? fields:

Code Name Description Category URL Public Database?
XI-PB Public Bodies - Open Knowledge Foundation Public Bodies is an Open Knowledge Foundation (OKF) project that aims to provide a unique ID for every part of every government. The main website is a portal for linking to profile pages for countries whose public bodies OKF has recorded. There are also links to directly download CSVs of this information. Data available on the following countries/regions - Brazil, European Union, Germany, Greece, New Zealand, Switzerland, United Kingdom, United States BR http://publicbodies.org/ False

Do you know how this could be fixed?

augusto-herrmann commented 2 years ago

@timgdavies and @rorads do you know how the process is on IATI, in order to fix the "public database" entry for PublicBodies from False to True? Because the data is, in fact publicly available.

I think this is the only think keeping us from closing this issue as finished.

bill-anderson commented 2 years ago

Email support@iatistandard.org

rorads commented 2 years ago

Hi @augusto-herrmann - I'm not completely sure what the process for this is now, as a lot has changed since I left IATI, but I think it might be as simple as submitting a pull request here flipping the 0 to 1. Might also be worth raising an issue on the repo to see if the IATI team are happy to do this quickly. (Edit: or as Bill says, email support!)

Parallel to this, org-id.guide still exists and lists Public Bodies here. I think you can follow these instructions, or raise an issue on this repo.

augusto-herrmann commented 2 years ago

Thanks, @bill-anderson.

I sent an email:

Hi,

I am one of the maintainers of the Public Bodies project, which gathers data about the structure of public organisations around the world. I have noticed that it appears on the IATI list here:

http://iatistandard.org/202/codelists/OrganisationRegistrationAgency/

and it is listed as having "Public Database" with a value of "False". However, the data is publicly available here:

https://github.com/okfn/publicbodies/tree/master/data

So I believe this should be displayed as "True". Could this information be rectified, please?

PetyaKangalova commented 2 years ago

@augusto-herrmann and all,

IATI no longer maintains this codelist. We only maintain XI-IATI codes. The process will be to identify if XI-PB is available in org-id.guide. I can see it is available (http://org-id.guide/list/XI-PB) and also see repo here and suggest a change by creating a pull request.

As of 17 July 2017 this list is maintained by the org-id.guide project. Data publishers can now search for and locate the relevant list for a particular organisation identifier using the org-id.guide website. The full register of identifier sources is also available to download in XML, JSON and CSV formats. IATI periodically replicates the codelist of identifier sources from org-id.guide, to assist those accessing IATI documentation. However, it is advised that the most up-to-date source is the org-id.guide project.

augusto-herrmann commented 2 years ago

Considering that the IATI codelist is no longer maintained (as highlighted by @PetyaKangalova above), and that we are already present at its successor list (org-id.guide), if no one has any objections (esp. @danfowler, @timgdavies & @rufuspollock who were involved when opening it), I'll soon close this issue as solved.