In attempting to find how a participating organization has been classified by other publishers, I just realized that the data store appears to only allow for searching on IATI publishers in the participating organization field.  However the organization I'm looking for and many other participating organizations are not IATI publishers. 

I can also see in d-portal that I can search on the organization name in the free-text search bar and find that the organization is listed in many IATI files but I don't see a way to get a dump from d-portal that can help me compare the org type across all those publishers.  

In this particular instance I'm looking to see how GIZ is classified by other publishers.  (government, public private partnership, CSO)  But I have others that I need to look up so I need a method to do this repetitively for other organizations.

Questions: 

1) Are there any data users out there with any advice on how to get to this data without going one-by-one into each activity spit out by D-Portal?

2) Are there any publishers out there that are willing to tell me how you have classified GIZ in your files?

3) Can someone from the IATI technical team explain why the data store search is restricted to publishers and is there is another way to get out all participating organizations regardless of publisher status? If there is no solutions I will advocate for a user case here. 

Comments (14)

IATI Technical Team

Dear Michelle,
You have raised very interesting use cases. We recognize that at the moment there is no easy way to search for non-publishers as participating organisations or to get a list of all participating organisations. As you have rightfully noticed, the Datastore lacks this functionality and d-portal doesn't allow you to download data in bulk to compare data across different publishers.

We can add a new user requirement request with the Datastore team to possibly include a free-text search for participating organisations that will search the participating_org_narrative field in the IATI Standard.

While at the moment there is no easy way to get this data, technical users can access this information using the Datastore Solr API (https://iatidatastore.iatistandard.org/documentation). For example, to search for "National Farmers Federation Moldova" non-publisher organisation name they can use the following call https://iatidatastore.iatistandard.org/search/activity?wt=csv&q=(partic…), which will return the list of activities in a CSV format.

Michelle Levesque

Well a HUGE thank you goes to Shi at DevInt for providing me with a data dump on GIZ. I've been able to discover the following about GIZ which I thought may be of interest to others.

1) Not surprisingly, there are 65 different versions of the name found in the files.

2) There are 9 different organization types used to classify GIZ. 10-Governement is the most prevalent (16 publishers use it) but 15- Other Public Sector, 21 - International NGO, 40 - Multilateral, 60 - Foundation, 70 - Private Sector (2nd most popular), 72 - Private Sector in Aid Recipient Country, 73 - Private Sector in Third Country and 90 - Other have all been used too.

3) There are 10 different organizational references used. DE-1 and XM-DAC-5-52 are tied as the most used with 4 publishers each. It is worth nothing that DE-1 (Germany) as a publisher uses XM-DAC-5-52 so perhaps we should all be going with their lead. Unfortunately they don't provide an org type classification for us to follow.

A few other things I noticed:
1) It seems some publishers have put their own org ref in the place of the participating org ref field. Others have put N/A and still others have put in numbers that don't look anything like an org ref.
2) With a search on GIZ (to ensure all derivative names are picked up) you get three participating orgs that are not GIZ so name search isn't fool proof.

I do hope the technical team is able to provide a better data mining tool for non-technical people sooner rather than later. Having the ability for publishers to look at other existing data as a guide to org references, org types and names used for participating orgs being added to their dataset for the first time would hopefully help with data quality/consistency. It sure beats guessing or potential hours of research to come up with something from scratch.

Justin Senn

Hi Michelle. In case it helps, here is an example of a contribution from GIZ to UNHCR: https://d-portal.org/q.html?aid=DE-1-201849652. I notice BMZ are both the publisher (with DE-1) and are also listed as "Funding" the contribution. GIZ (with XM-DAC-5-52) have the roles of "Accountable" and "Extending". UNHCR publish transactions from GIZ with DE-1 as the provider-org. The IATI standard tells us the provider "is the organisation from which the transaction originated". Do you know if there is a clear definition of "originated" for this case? I guess it could be either the organisation that is publishing, funding or accountable. Please excuse my ignorance if there is an answer already shared.

Michelle Levesque

Justin,

I don't believe IATI has gotten that detailed in its guidance but IOM has always taken the position that our funding organization in IATI is the one in which we have contracted in the grant agreements. If the donor received those funds from another place we don't include it in our files. In fact our current ERP system doesn't capture it in a way that we could include that level of upstream funding info in IATI.

The challenge we have with DE-1 publishing on behalf of GIZ (listed them as accountable and extending) is that their traceability references probably don't line up with ours. I haven't done that much research again because I struggle to extract data on my own but unless what GIZ has in their contract as a reference is consistent with what BMZ uses in theirs I would venture to say there is a data gap in the funding chain. Research for another day......

And I appreciate you helping provide insight. Never any ignorance. Au contraire I still owe you a fondue.

Evgenia Tyurina

Dear colleagues, thank you for this very interesting discussion. I thought the main source of organization identifiers was the IATI Registry (https://www.iatiregistry.org/publisher ). There are only two publishers for German Government in IATI Registry: "Germany - Federal Foreign Office" (XM-DAC-5-7) and "Germany - Ministry for Economic Cooperation and Development" (DE-1). We contacted the Ministry for Economic Cooperation and Development to confirm the coverage of their identifier and they confirmed that DE-1 also covers GIZ and KfW Development Bank, so we were going to use DE-1 whenever we receive funds from these two donors.
Would you know where the XM-DAC-5-52 ref for GIZ come from?

Amy Silcock

Hi all, the GIZ code XM-DAC-5-52 comes from their OECD DAC's agency code. You can see this in the 'Agency' tab in the attached DAC CRS Code file. Also accessible here: https://www.oecd.org/development/financing-sustainable-development/deve…

There's a comment in the draft traceability guidance about what publishers 'could' do when the IATI publisher reporting the funding to your activity, is different to the organisation you recognise as the funder. E.g. DE-1 reports rather than XM-DAC-5-52. This is up for discussion! Please do raise as a question/views in the feedback form: https://iaticonnect.org/discussion/consultation-help-shape-iati-standar…

Evgenia Tyurina

Hello Amy,
We (ILO) looked at the draft guidance and will send our feedback in the feedback form by the indicated deadline. However, it looks like the draft guidance already provide a clear answer for the above: "For incoming transactions the reporting org should list the provider organisation/s it is receiving funding from. You should use the name and IATI Org ID the provider organisation publishes under, even if this is different to how you refer to them. For example, you may receive funding from a specific ministry or department, but your provider publishes as an entire organisation or government." We understand from this para that in the GIZ case we should use the "IATI Org ID the provider organisation publishes under" which is DE-1.

Amy Silcock

Thanks Evgenia, we added it in based on this Connect thread. It's a new piece of suggested guidance :) Very much open for discussion and ideas of how this should be handled!

matmaxgeds

Hi - as I understand it, GIZ is legally a private company, and so 70 - Private Sector (or 73 where implementing a non-german-funded project) seems most appropriate? Or does IATI anticipate that differnt codes are a good thing as different organisations have different perspectives? Having different codes in use might make it very difficult e.g. as a researcher to use the data (you cannot filter by the org_type). If so, would it be difficult to notify all the publishers who are using different codes to get them to update their data? Maybe this should be a regular exercise in cleaning the data - who would be in charge of that kind of data quality?

Michelle Levesque

Max - I would highly recommend we not all select different codes and types only different roles if appropriate. If we look at Amy's comment about GIZ having a DAC code it would lead one to believe it is part of the German government. The fact that the German Foreign Affairs ministry publishes their data within their info further clouds the issue. Like you, I found GIZ to be a private company with a registration as a private business and so we initially used their private registration info as their identifier and called them private. We now use their DAC agency code and we list it as a government. This is precisely why I brought up the topic. It is as clear as mud and lack of ability to search and dump across all participating orgs (not just publishers) further complicates matters. As a publisher I'd like to be able to compare and contrast participating orgs from multiple perspectives. (roles, types, org ID). We'll get there but for the moment I struggle.

matmaxgeds

Hi (Matt Geddes here) totally agreed - I guess the key question is what is the process for resolving this, as it really appears to be a key service for IATI to offer publishers and users to improve data quality e.g. extract a list of all the orgs with multiple roles recorded, and to contact the organisation, get them to choose, and then share that information with everyone. I don't think it can be worked out with logic via the 'rules' as we are all finding, as there are too many exceptions to each case.


Please log in or sign up to comment.