Since early 2016, we are helping organisations produce and improve their IATI data. We created tools to visualise their data and show what the data is about. And we started delivering automated data quality feedback, based on our experiences with “data in the wild”.
In collaboration with both the Dutch Ministry of Foreign Affairs and the DFID in the UK, this data quality feedback tool has evolved into an online IATI Data Quality Feedback platform that processes all public IATI data and makes the feedback available to everyone, creating a true open data source.
A quick tour
It may be a little overwhelming at first: for each activity, there usually are multiple messages on how to improve the data. For a large dataset, the list may look daunting. There are lots of details to get right to make use of the full potential that IATI has to offer, but not everything is essential for each case.
Just look around for particular areas of improvement: often, there are simple fixes that can improve the data and reduce the number of messages, or that indicate an error in the system producing the IATI file.
Most feedback is based on the IATI standard and common practice in the IATI community. There are also additional rules for donors, for example a requirement to include OECD DAC sector codes in the classification of an activity.
We divide the feedback into severities:
- Errors make it hard or even impossible to use the data in a meaningful way.
- Warnings give guidance on how to increase the value of the data in a broad sense.
- Improvements are ways to still add a more value for more specific cases.
- Notifications may or may not point to problems, or may indicate ways to optimise the data.
We also divide the messages into categories. These correspond with the general setup of IATI. Specifically, users of Aidstream should be able to recognise the section within AidStream where they need to go to make changes.
How did we get here?
IATI data is written in a technical XML standard, and it is possible to validate your data against this XML standard, but this leaves you with two problems:
- The error messages refer to line numbers in the XML file. These are useful for programmers, but hard to interpret for staff members. It is also hard to link back to the actual source: in which activity and in which area of the data is the error? You have to know how to read the XML code.
- The validation only covers a minimal technical compliance. It does not check the additional guidelines by the IATI community or the requirements of donors. So it does not include tests like: is the start date of your activity before the end date? Did you include an “incoming commitment” from the donor?
We needed a better way to help organisations improve their IATI data, and so we developed the Quality Feedback tool. In our work with dozens of organisations, we used the feedback in workshops and online, to give more actionable advice to publishers.
We have already made several changes based on feedback from those users, and we have plans to further develop the platform. For now, we invite you to have a look at it too, and to let us know how it helps you improve the quality of your data!
The Data Quality Feedback platform is part of our vision on IATI data. We want to help organisations produce and use IATI data in meaningful ways.
- We have a Spreadsheets2IATI service, to create IATI files from simple-to-use spreadsheets.
- We build Dashboards to help programme managers get insights into the combined IATI data of activities by published by multiple organisations.
- We do data science research into IATI data to explore networks and potential opportunities.