Summary

IATI Data Bugtracker is a public log of issues with IATI data (the information contained in IATI datasets, about activities and organisations) and metadata (the information contained on the registry about the datasets and publishers). It makes these issues more transparent by making them public and shared. Anyone can create an issue, or comment on an existing one – you just need a free GitHub account.

Issues are organised according to the publisher, so it should be easy to see related or repeated issues. Issues are also rechecked every 30 days, to ensure they’re kept up-to-date and don’t become stale.

Rationale

It is widely acknowledged that data quality issues exist within IATI data, and that these issues can be a barrier to data use. They need to be addressed by publishers, in order to make the data more reliable and usable. The IATI Validator is a step forward in tackling data quality issues, through automatic schema and ruleset validation.

Many data issues are more nuanced, and are therefore harder for the Validator to spot. Publish What You Fund’s Aid Transparency Index attempts to publicly address data quality issues by manually reviewing IATI data for some of the biggest publishers. Within the IATI Secretariat, the Technical Team’s Business Analysts also take action to address data quality issues. This is not done publicly, though – it’s conducted between the secretariat and the publisher. It seems likely that a transparent process may be more effective in getting issues resolved. More people can suggest solutions, and the problem of several groups working separately on the same issue (with potentially different approaches) is reduced.

How it works

IATI Data Bugtracker is a minimum viable product. We’ve tried to do as little work as possible to get something off the ground, to test whether the concept is workable and useful. Here’s roughly how it works:

  • We have templates in place for different types of issue. These help us get clear reports from data users about the problems they’re experiencing.
  • We ask the data user to also contact the publisher directly and let them know about the problem. We’ve provided an email template for doing so, and we ask the user to confirm that they have indeed contacted the publisher.
  • Issues are automatically labelled with the publisher’s name (so we can easily see a list of all issues that relate to a particular publisher)
  • After 30 days of inactivity, a bot will check in, to ask if the issue is still valid.
  • If someone replies to say the issue is still valid, it’s automatically moved to “Long-standing issues”. If the issue is instead resolved, it’s moved to “Fixed!”
  • There’s a project board, for an overview of the status of all issues

Issues are also shown on Code for IATI Analytics (we’ll talk more about this in a later blog.) E.g. on the Australia DFAT publisher page. Only live issues are shown – resolved issues are removed.

Results so far

IATI Data Bugtracker was created in April 2020, as a way to maintain a public log of data issues. Since then, it has successfully played a part in resolving several issues. Some examples:

A special thanks again to @notshi from d-portal, who is the number one contributor, and has successfully helped facilitate several successful resolutions.

Please take a look at the project board if you’d like more details on how it has worked so far, and by all means please contribute with any issues you spot.