This Saturday is open data day. I’ve been working around the field of open and data (separately and together) for ten years now and I have been thinking about what next for open data as a concept and what both providers and consumers need to give the concept a new lease of life.

It’s worth pointing out here that plenty of open projects that produce open data are doing very well and especially projects like WikiData have picked up a lot of the slack of creating infrastructure and caretakers for data where previously there was none.

That said, here are my suggestions for what should happen next.

Sometimes things just need to be open-er.

I’m a great believer in the old GDS sticker “make things open, it makes them better”. I think that the revolution in insisting that lots of parts of government need to be open by default and closed by exception is where things should start from. However. I’ve worked on several projects in the past few years where I’ve been concerned that a completely open dataset of seemingly innocuous data would create unacceptable risks. This isn’t to say that the data should be closed, but that an audit log of responsible use of delicate, but public, data is in some cases sensible and often it seems hard to find the language to talk about things that should be “trust but verify”. We cannot live in ignorance that data, even without personally identifiable information, can be socially harmful and in civil society and civic technology, we should be thinking in terms of consequence scanning and “red team” exercises to play through what the worth that could happen is and thinking about what can be put in place to safeguard against that.

I'm interested in the work that's ongoing around data trusts, but sceptical that that model of governance is going to fit all circumstances, and I think more exploration of the grey areas on the edge of open licensing is vitally needed.

Too many excuses

That said, those cases are really rare and unusual. More often, over the past ten years we have seen the public sector reverse course on many of the promises of earlier administrations. FOI compliance from Whitehall has dropped dramatically. Open data releases on are often late and unusable. The Government Digital Service was so unable to make its (flawed) performance platform work that they canned it and introduced more barriers to citizen access to data on how government is doing. These all seem to have occurred as the standard of government has got objectively worse. The Open Government Partnership recently placed the UK in procedural review for repeated infringement of practices.

Government is not doing well enough, and that isn’t just politicians, it is civil servants. The air of needing secrecy is a security blanket to protect against legitimate critique. Those involved in the practice of government need to see open data as a good faith promise of honesty and the embodiment of upholding the civil service code. Politicians, similarly, need to be held to standards: plenty of standards exist, but the mechanisms for enforcement seem to be in abeyance.

Data user needs

User needs are a slightly/massively reductive way of viewing the world, but in terms of open data, the “so that” is always missing. Armchair auditors never appeared in the wake of open data release and civil society’s oft repeated call of “holding government to account” rings hollow as there basically isn’t any mechanism to do this. Making yet another website that shows the government is performing like shit is having no impact. So, as a community, we need to articulate what we need data for and how we can use it for purposes that are emancipatory and impactful, rather than just another map.

I want to keep fighting for open things in the UK. I hope you do too, and I hope we can start to take forward ideas that have a bigger political imagination that matches the challenges we are facing.

Happy open data day.