“That was much more interesting than I expected!”
I got this response after leading a discussion with information managers and archivists on the work to create an authoritative dataset of government organisations, otherwise known as Commitment 11 in the Open Government Partnership’s National Action Plan.
Since the draft open standard was put together (thanks, Cam!) we’ve been continuing to talk to people working inside and outside of government who own the data, use the data or are interested in describing the ‘machinery of government’ (the organisations, roles, and process of government) in a consistent way.
With the help of the great open data people at Stats NZ we ran a meet-up open to test the data model and check the scope.
There was also a robust discussion about the OGP National Action Plan at NetHui this year where Commitment 11 was discussed with interest and excitement. It’s brilliant that so many different people recognise the value to government accountability and transparency this work could bring.
I am continually grateful for the support of people in civil society who give up their time to share their knowledge and experience to make this work better (looking at you Andrew Ecclestone and Laurence Millar). It is a major benefit of having work included as an Open Government Partnership. Collaborative working with interested parties outside of government, in my experience, makes for a better outcome.
Feedback from the open government and open data communities
There are some areas of consensus on the data model and open standard, for example:
- Unique ID needs to be a running number, not Legal Title.
- Multiple smaller tables of different fields are better than one big flat file.
- Needs to be extensible, agile, flexible to accommodate future structural change.
Additional fields for the data model have also been suggested:
- Needs to include superseding or successor agencies.
- Capture actions e.g. merge, demerge and why this occurred.
- Capturing relationships with satellites and subsidiary organisations – types of relationships.
The whole set of notes for the open data meetup are available for anyone who wants to dig deeper.
Scope of the dataset
What this dataset could cover has been talked about a lot in the meet-ups. There are so many great use cases - from historical tracing of agencies for accountability, to the efficiency gains from an up-to-date list of contact details could bring to government - which make the discussion of scope hard to wrangle.
The focus on an extensible model has multiple benefits. Being able to extend the model means that it can be used for many more things, and there is also less pressure to get this perfect.
However, it is clear that while the idea of a ‘skinny’ set of data about government organisations is a good place to start, we would not achieve the desired benefits of this OGP commitment if the work stopped there.
The ‘skinny’ dataset would focus on data that doesn’t change too frequently, such as government agency name, type of agency (e.g. Public Service Department, Autonomous Crown Entity), and Ministerial portfolio. This is in keeping with the work happening in Australia called the Longitudinal Spine of Government Functions (LongSpine).
An aim of LongSpine is to allow changes in structure and function to be tracked over time.
This makes information about government functions and structure more readily available and improves the integration of other sources of information on topics such as staffing, budgets and reporting.
We’re keeping in touch with this work with the help of Archives NZ who are starting some amazing related work on linked data.
An example of this is the requirement for survivors of historical abuse in state care to find out which government agency holds information about them. The information they’re looking for spans the time period 1950-1999.
This is a long time in the life cycle of government agencies and requires survivors to know the ‘lineage’ of government agencies. There are several government organisations who created records on children in state care and their names change over time.
Ministry of Social Development, for example, has been known as the Department of Social Welfare. It’s intervention and care of children has been called the Department of Child, Youth and Family Services, Child, Youth and Family and is now known as Oranga Tamariki. This knowledge is really difficult for survivors to find out.
We’ve also been excited to learn about the work on organisation ontology done in the UK, since getting the naming right is crucial to the success of this piece of work. Learning from, and building on, the work done by other countries is a key idea of the OGP, which often says that ‘good ideas come from everywhere’.
We’ve got a group of government officials and civil society people who meet to continue this work. Our last meeting focused on exploring the data model behind the Te Kawa Mataaho Public Service Commission’s list of state sector organisations.
SSC will be releasing some of this data on data.govt.nz early next year, which is good news. It’s great to have their contribution to this, as the value of this work will come from it being authoritative – meaning agencies use and maintain the data themselves for their own needs.
Next up the group will be looking more deeply into modelling government organisational change e.g. mergers, re-namings, disestablishments etc.
Do get in touch if this is something you’d like to find out more about or if you want to be involved in this work ongoing.
One thing is clear, the work is an essential foundation, not only to support a wide range of information and services about government, but also to enable its digital transformation.
Get in touch at firstname.lastname@example.org.