OKFest - Humanitarian Sprint

An idea has surfaced to also host a Humanitarian Sprint on the OKFest fringe event.

This could take the shape of:

  1. Working with some existing 3W field datasets (OCHA and clusters/NGOs), looking as a group at the lalignments and
    misalignments between that and IATI to consider recommendations for both IATI and HXL.
  2. Looking at the draft HXL standards and their alignment with IATI.
  3. Reviewing any existing huimanitarian work going on with FTS.

We’ll continue to look into this as an opportunity and gather idea here.

If you’re interested in hearing more or contributing to/joining this session, please post your thoughts below.

Hi, Adrian, and thanks for starting the discussion.

Humanitarian data has been a big topic for IATI for the past year and a bit, and I thought the OKFest side event could be a good place for us to dive into some of the practical details from a developer perspective. For anyone new to the topic, humanitarian aid involves short-/medium-term responses to crises like last year’s Philippines typhoon, while development aid (IATI’s main focus to date) involves helping to build long-term social, economic, political etc. infrastructure in partner countries

Existing humanitarian data already has a lot in common with IATI’s standards — for example, the humanitarian community collects “3W” (who-what-where) data for nearly every crisis, which is very similar to IATI Activity data — but it’s also different because of its scope, emphases, and timelines (for example, needs are usually more important than finances). We’ve talked about these issues at the policy/high-tech level, but I think it’s time now for us to dive in and look at what humanitarian IATI data would mean to us as software and data specialists.

I think it would be a great idea to spend some time on the Friday looking at this. @David_Megginson - really agree re diving in (and the “needs” suggestion is also very interesting - also for development aid?). I think if we can come up with rough bare bones of how IATI might need to adapt to better respond to humanitarian needs then it might kick-start some other conversations.

Talking about humanitarian data brings up the bottom-up discussion that’s relevant for development data as well. Right now most (not all) IATI data is top-down — we’re getting the cooked, agreed-upon, internally-consistent(ish) data from the various HQs. We get something similar for IATI by going through FTS.

In humanitarian 3W data, though, there’s a big emphasis on the raw, imperfect, inconsistent (but fast-delivered) bottom-up data, e.g. what do NGOs actually report that they’re doing on the ground, vs what donors think they’re doing. This data is rarely financial — it’s more for forming a common operational picture than for tracking money — but it includes many other aspects of an IATI activity report. The challenges dealing with this data in the humanitarian world will be similar to the challenges we’ll face dealing with it in the development world.

I’ve been analysing humanitarian 3W templates and spreadsheets, and I’ve found that these are the top 11 fields so far (including the corresponding HXL codes):

  1. adm1 (admin level 1 description): 100%
  2. adm2 (admin level 2 description): 100%
  3. sector (sector/cluster description): 100%
  4. impl (implementing org): 85%
  5. adm3 (admin level 3 description): 69%
  6. prog (programming org): 69%
  7. status (project/activity status): 69%
  8. subsector (subsector description): 69%
  9. from.date (project/activity start date): 62%
  10. to.date (project/activity end date): 62%
  11. targeted.num (number of beneficiaries targeted): 62%

As I mentioned on the HXL mailing list, this suggests that a minimum viable humanitarian 3W report could include just three fields: adm1, adm2, and sector. In fact, that is enough to draw a basic activity map (or one for each sector), and generally, to form a very high-level common operational picture of a crisis response.

So when we come back to the IATI side, it’s worth asking whether the “T” for transparency always stands for financial transparency, or whether it could stand for other types (e.g. transparency about what the aid community is doing and where). Fewer than 25% of the samples I’ve seen contain budget information on an activity level, and none show expenditures. Is IATI’s tent big enough to include raw field activity data like this?

One more subtopic: how should people and orgs doing IATI monitoring/scorecards (like PublishWhatYouFund) weight data completeness for humanitarian report? Some of the fields that are important for development aren’t important for humanitarian, and vice-versa. Perhaps we could work up a proposal as part of the sprint.

Summary video from the IATI Humanitarian Sprint - https://www.youtube.com/watch?v=qul_pDcanzY&feature=youtu.be

Great context

So what to do?

  1. Get a humanitarian marker in place
    2, Bring FTS extension covering emergencies, appeals, etc, into the main standard
  2. Redefine timeliness for humanitarian needs: best practice is daily, not monthly
  3. Focus on Result Outputs: Get WFP in the saddle for this
  4. Get into a situation asap that EDRIS and FTS feel comfortable with importing IATI data