Enabling Two-Way Traceability

(Wendy Rogers) #1

As has been pointed out by @Michelle_IOM in another post it is already possible for donor and recipients to cross reference each other and the IATI Standard does already make provision for that to happen. However, as @Herman had mentioned (and it was also my understanding) I think that the original intention for traceability within IATI was just to trace upwards within the funding chain? Both @Herman and @YohannaLoucheur have pointed out that a donor or funder can generally provide their own originating IATI activity identifier to the recipient via contract documentation and/ or as part of the contract management business process etc .

Interestingly, some of our most recent work within the Grand Bargain (particularly around issues relating to Localisation) is highlighting that there would indeed be value in being to be able to trace both ‘up’ and ‘down’ the funding chain. Presumably for this it happen, Donors and other funders would need to regularly ‘harvest’ the recipient activity identifiers and add them into their own systems. I think this process could certainly be automated and it would also provide donors with an automatic confirmation or validation that recipients have also now published the receipt of that funding to IATI? As a result I would be very interested to hear any views from donors or others on the practicalities and appetite (or lack of?) for making this happen?

1 Like
Using related-activity to link data between different publishers
(Andy Lulham) #2

This is great Wendy – thanks!

This seems like a good opportunity to post this visualisation we made last week:

This shows a snapshot of all links, both up and down the funding chain. The links shown are reported via the attributes transaction/provider-org/@provider-activity-id (i.e. “upstream”, shown in black) and transaction/receiver-org/@receiver-activity-id (i.e. “downstream”, shown in red).

It’s totally possible to infer the link from provider to receiver when the reciprocal link is provided, and vice versa. What I mean by this is…

  • Let’s say an implementer publishes an “incoming fund” transaction.
  • This transaction links to the funding activity, using the @provider-activity-id attribute. That’s an upstream link.
  • It’s totally possible for a tool like d-portal to then show a link from the funding activity down to the implementing activity, even if the funder doesn’t provide the downstream link.
  • It might be desirable for the tool to label this as an “unverified link”, since it hasn’t been confirmed by the funder via a downstream link. But it’s absolutely possible to show these links, as the visualisation above demonstrates.

I agree that showing these reciprocal links would be A DESIRABLE THING! Agree also that “verifying” them in provider data (by publishing the reciprocal link) would also be great.

(Matt Geddes) #3

What would happen if we had an ‘upstream link’ bot operating as a secondary publisher that provided all these upstream links - that way they would be available to all data users whose systems can cope with information on one activity coming from multiple publishers? The ‘unofficial’ nature could be stated in the bot org file.

(Wendy Rogers) #4

Thank you both @andylolz and @matmaxgeds and it’s good to have a confirmation that this is possible. Also @matmaxgeds I’m interested in your suggestion of using a bot but (and I am probably missing something obvious here) I am just try to work out what its added value would be as opposed to any data user interrogating the IATI dataset directly to identify any related links? Would you be able to elaborate TIA

(Matt Geddes) #5

hi @Wendy - nothing obvious to miss, my suggestion of a bot was designed for the situation (as common now) where data users are not making time to interrogate the IATI dataset to identify related links - that remains the best solution, but the bot could fill the gaps where this isn’t being done, so that we don’t miss out on the the improved data quality this technique allows because of a bottleneck on the publisher end.

I think one of the great strengths of IATI that is a bit underused at the moment is that other/secondary publishers can augment the data provided by others (e.g. also for machine translations) and link it to the original activity through the activity-code - I thought that this might be a good opportunity for that - it would also mean that bu doing it once centrally, it might be much lower cost/effort that us all having to work them/develop tools individually.

(Wendy Rogers) #6

Thanks for the clarification re the use of the bot @matmaxgeds . Really helpful and makes sense.

Its also interesting that you raise the issue of secondary publishing here. As part of our work on the Grand Bargain we are just starting to look at (with our other hum data provider colleagues such as for FTS and HDX) future options for data provision within the humanitarian sector and ‘improved’ secondary publishing could have a significant role to play? Whilst we are only in the very early stages of our thinking it is something we hope to be able spend more time on in the second half of 2019 (other priorities permitting). As a result I hope to be creating another separate post at that time to discuss the potentially enhanced role of secondary publishing with the wider IATI community.

(Herman van Loon) #7

Basically the process @andylolz describes, is deriving the downstream link. This is i.m.o. really an data use issue and not an IATI publication issue.

Publishing both links, increases the chance of introducing inconsistencies. Another reason against publishing both, is described above: at the moment the funder publishes the funding activity, the corresponding IATI identifier of the funded activity is usually not known.

So if you need the downstream link, derive it and use it in your application. But do not require to publish redundant data.

As far as the point @matmaxgeds makes about secondary publishers: I would say this is one of the weaknesses and not strengths of IATI since secondary publishers obfuscate data ownership and introduce the same (but slightly altered?) IATI data of a primary publisher twice in the IATI ecosystem. What is the truth and who is accountable for data quality in that case?

(Matt Geddes) #8

@Herman - publishing an downstream link mirroring a upstream link wouldn’t duplicate or alter any data as far as I can tell, but…

  • RE accountability for data - unclear how this would be any worse than for primary publishers - IATI doesn’t even require an email address to publish?
  • RE data appearing twice - The same projects are already re-reported many times down the implementation chain - multiple reporting is fundamental to IATI’s approach.