An Open Data Camp 7 session on countering excuses for not publishing open data, led by Jenny Broker. Liveblogging: prone to error, inaccuracy and howling crimes against grammar and syntax. Post will be improved in the coming days.
Excuse: It’s a safety thing – it’s critical and it could be useful to terrorism
Safety is the first thing people will come after you with. For example, in utilities, it’s a very real concern, particularly around the location of assets. Is this a genuine concern, or an easy way of shutting down a conversation? Is this information that’s not already accessible via Google Maps, for example? Crashing critical infrastructure is a genuine risk. The most risky data is already heavily controlled — and is often not even shared within government. That comes with its own problems – issues get missed because staff don’t have access to the full picture.
So, if Google Maps has the data, if we make it more accessible, is there a potential for spotting problems earlier? Well, liability now raises its head. Pretty much all datasets are infested with personal data, so if you published the data, and something happens, you’re liable. Some people don’t want to take that risk. This is another standard way of hiding from open data. Some organisations have developed organised risk assessments for open data – it create a more structured way to talk about risk.
But it’s also worth checking if anyone else has already published the particular dataset you’re after…
The Data Ethics frameworks arose to counter some of these excuses. Can you start to make the case that there are benefits that might outweigh the risks? Could local responders react more quickly to an emergency if they had easier access to the data? There are certainly opportunities for efficiencies. This is akin to the open source argument of “many eyes”.
There are language issues – people talk about things in different, incompatible ways, and until there’s communication at a senior level this is hard to change.
There’s also plenty to learn from the Defence industry and establishment, which are very good at knowing what data can be released – and which must never, ever be released.
Excuse: Why would I do this? It’s just more work for me
The fear of personal liability – or a burden of work – is common. There’s a tension between the greater good and the practical. It might not be in your job description – or your interest to open data. You might just lack the time. This could be addressed by a central data owner, well-trained, with a familiarity with issues like GDPR. They would almost be like a policy team for data. Or a data parent, perhaps.
This would help address governance issues. It’s someone to assess use cases — and use cases are one of the best arguments for getting data opened. Government departments often run a “call for evidence”, where these cases might be made.
There also needs to be more demand — “I need you to do this so I can do my job” — because once the data is being created, it might as well be put out there. Prototyping to demonstrate the benefits can get other people enthused, and help drive demand. Publishing lists of datasets held can also help stimulate that demand.
If the data quality isn’t perfect, disclaim that. Don’t be embarrassed by it. Get it out there, and then you get the symbiotic relationship with the consumers that shows you where you can improve it. Send the data out with metadata, which is explains why it was collected, how it was collected, and the limitations.
Open the right data, partially open other datasets – and never share a third group. It’s about sharing appropriately. There’s no sense that the ideal is open data. There will always be data that can’t be released.