A guest post from Bill Anderson of Development Initiatives on the challenge ahead to make sure we have the data that matters.
Decision-makers and those that hold them to account, whether at local, national or global level, require access to usable, meaningful information that throws light on the problems they are seeking to solve. Putting information into context invariably requires joining up data from a variety of sources.
In the past five years open data champions have had many successes persuading both governments and others to open up and there are a now hundreds of public portals containing a huge variety of datasets. Yet most of these repositories are silos: containing similar data, yet in different formats, structures and standards.
It remains incredibly difficult, for instance, to assess government spending in the context of the demographics and impacts of its intended purpose. It remains a huge challenge, often with life threatening consequences, for humanitarian aid workers to match financial and material resources against needs and conditions in the midst of an emergency. We clutch at jigsaw pieces and very rarely see the finished picture.
One of the biggest blocks to joining up data is a lack of compatible standards. We are faced with different rules for access, different data formats, different data definitions and different quality guidelines. This poses a number of challenges.
- Technical standards need to streamline and simplify the way in which machines are able to share data – both with humans and other machines
- Publishing standards need to ensure that similar concepts are both comparable and interoperable across different schema.
- Global data standards defining such things as geospatial and functional entities need to describe the world in a common, or at least translatable, language
- Data quality standards need to ensure that data is not only logically valid, but that it is both accurate and meaningful.
There is, too, a bigger standards picture: often the data presented to us is not what we think it is. Here’s an example.
The number of women who are still dying during childbirth is an obscenity. Reduction of maternal mortality was a key target of the Millennium Development Goals and remains one in the post-2015 SDGs. How do we know how many women are dying? In 65 countries they are counted through the national registry of deaths. In the remaining 116 countries maternal mortality is calculated through an algorithm using the GDP, a (calculated) fertility rate and an (estimated) statistic called “Skilled attendant at birth”. In other words we have absolutely no idea how many women die in childbirth. This is why the call for a data revolution for sustainable development is gathering steam with African countries in the forefront of the struggle.
We have serious work to do in Ottawa. Data users, infomediaries, decision-makers, watchdogs and others frustrated by a lack of meaningful joined up data need to articulate their demands and put pressure on all standards bodies to get their acts together.
What are your priorities for action on open data standards? Share your thoughts in the comment thread below…