Everyone agrees we should have it, but in the busy worlds we operate in it’s the first thing labelled as “good to have” but not essential.
This is dangerous as government departments gather economy-driven information, which, in order to be managed properly must be accompanied by comprehensive, machine-readable metadata.
This is not an easy task.
Different departments have different data estates, some are technically more mature, and others cannot fully invest in their metadata solutions just yet. However, it is still critical that metadata should be associated with data.
Getting serious with metadata
If we are serious about metadata and encouraging others to take it seriously then we need to do more to make it easy for departments to do that.
Across government there is more sharing of best practice, developing guidelines and establishing a “development strategy” for departments to become more data (and hence metadata) driven institutions.
This is not a trivial task and it requires both a comprehensive understanding of the metadata needs in a given department and a policy to ensure cross-government consistency in gathering it. This is paramount in making sure that we can publish and share data supported by metadata with each other. Recently, the Government Digital Service (GDS) have proposed a set of data tags to help other government departments which might be struggling to help them collect metadata, when none is available.
Sounds good in theory? It is a lot more complicated than that. While what the GDS have proposed is incredibly useful, it should be the start of the process and be expanded upon and developed further.
Indeed, the standards to which government publishes its metadata across departments should be standardised. This would ensure that data does not lose its meaning and would be discoverable and accessible to people who translate it into insights and publications like blogs, reports or policies. This is useful work and departments should consider going further, if possible.
To boldly go…
We’ve tried to put that into practice ourselves at the ONS. We use elements in our Logical Metadata Model based on, amongst other international common standards, the Dublin Core-based tags as they are the building blocks of majority of metadata standards that exist. By adopting these principles, of reusing and adopting international, well-established standards, we ensure consistency and will be able to share comprehensive, common metadata with our data stakeholders. The future tense in that last sentence is there for a reason; we are currently testing our Metadata Model in practice through a Proof of Concept across our data estate.
Solving the problem now, solves much more later
A “minimal set of metadata” solution for government should be as simple as possible and should not include a large (or comprehensive) number of elements (i.e. tags). There is a danger that, if an elaborate Metadata solution with 30-100 tags were available, users would continue to use the provided minimal solution instead of taking the opportunity to implement more advanced standards, missing the point of a true transformation where metadata is embedded as part of a data management framework at the core of the organisation’s data strategy (which we are currently undertaking in ONS).
There are a few initiatives currently active which deal with the problem of producing consistent and standardised data and metadata across the government. These groups are key to seek consensus and give the opportunity for data leaders and practitioners across government to feed in their requirements. Some examples include the groups led / coordinated by the Data Policy, Strategy and Ethics at DCMS and ONS’s led cross-government Data Architecture Community (GDAC), although other consolidated groups exist that include data professionals in government.
The long and short of it is this, metadata can provide a world of opportunities to your organisation.
It holds the key to unlocking the true power of your data. It will allow you to share and use data more effectively without losing the meaning behind it.
It is not an easy task but looking across the sector and seeing the work already underway I am encouraged that we are taking the, collective, steps in the right direction.
We are addressing the issue of standards and policy well and the shy optimist in me dares to hope that together we can re-label metadata to “essential” and work on practical, extensive, requirements-based solutions to address our common problems.