Weeknotes S03E09
This week ended well. Some crappy isolated things happened. I feel far more in control than I have done recently. The week is brought to you by the words bridge and mesh.
Some things I’ve been doing and thinking about this week.
Data Platforms, Standards and Communities
I typically read quite a bit of short-form stuff each week and listen in on as much as I can in terms of events. I love the Institute for Government DataBites series. The session this week in particular. If you haven’t seen one, you’re missing out. The aim is to showcase interesting government data projects. The format is four quickfire presentations of eight minutes each, followed by eight minutes of questions from the audience. This week’s replay can be found here. They were all great but two relevant of comment here in the context of my wider week and work.
David R talked about the Data infrastructure of the Ministry of Justice Analytical Platform. Which prompted this tweet from another data David. David R rightly explained that different contexts need different technical considerations and have different needs.
This is an important point. The Ministry of Justice Analytics Platform supports their analysts. It needs to give them the right tools and access to data to do analysis. Then to get the results to whoever needs them in whatever form that might be. To do this the platform needs to play nicely with exisiting or new service data infrastructure to get the data. It may also need to play nicely with the infrastructure the users of analysis to get outputs to them.
Additionally if a platform needs to support code deployment into operational settings and services — like the one we are building in Health Digital — there are additional considerations. This can result in some context specific technology choices. So duplication of effort and money — as far as infrastructure is concerned — is an over simplification (see also the link to the distributed data mesh article at end of these notes).
The infrastructure we are building in DWP Health Digital is infra as code, pipelines as code to make it reusable/shareable and easy to add/swap tooling as we iterate. This could save duplication of effort and money in our context. Although even within and across DWP delivery contexts can differ.
The important thing is to make sure we can join up with elsewhere. Which is to do with standards and approach of the stuff that flows around the infrastructure as well as the end points. A big part of our foundational work is around ensuring data processes, standards and practices are defined and adopted within Health Digital. I mentioned in S03E07 that Daniel has been engaging with the Data Standards Authority. In many respects we are ahead of the game here. Rosalie from Data Standards Authority spoke at DataBites this week. Another great talk and I really like the direction of travel.
I spoke with Robin from Ministry of Justice at the beginning of their Analytical Platform journey four years ago. Stephen has linked up with David along the way. But this has been largely through our own efforts rather than any central co-ordination. I have mentioned the importance of strong communities many times in these weeknotes. I‘m hopeful the Data Standards Authority will make this community approach easier.
Centralised data strategy, decentralised data leadership
Speaking of Robin. He wrote a blogpost this weekend about command and control leadership in data. It covers a number of the elements I’ve touched upon over this most recent season of weeknotes. Right from week one (most strategies are too complex and abstract) and week two (building blocks and starting with user needs) through to week seven (bridging delivery in Health Digital and the central data part of the department) and week eight (culture and changing together). It resonates.
Robin describes a recipe for success which is a distributed model. What he decribes is similar to what I am attempting to put in place in Health Digital. Distributed data expertise and a focus on user-centred data products. This brings associated challenges and changes to how people think about data ownership and pipelines. Basically the operating model for data product delivery within a business area.
I recognise what Robin describes. He says the key elements for the distributed model are
- Understanding value and prioritising work.
- Empowering teams to solve problems.
- Coordinating teams and driving adoption of good ideas.
My role is bridging top-down centralised data vision and bottom-up distributed business-based user needs with strategic priorities . Turning talk into strategy into action. This week I’ve continued the conversations which echo elements of Robin’s post. I have been attempting to draw out my role. So that all areas are clear and it can be discussed.
In the spirit of being open, this is where I’ve got to this week. I am the blue square. The words in the curly brackets describe what I am currently doing. It hasn’t yet been agreed that it’ll continue in this way. I’m sure lots of this slide won’t make sense in isolation, its not designed for that really. But the gist should be obvious and it might be useful to someone. It is a work in progress. The numbers of teams/people/roles are illustrative.
Things take time
Certain things I’ve been hoping for, for quite some time, are starting to happen. From a centralised data persepctive and from a business area perspective. Also within our team. They are happening slowly but they are happening. During this time its been clear that having a team of data and design professionals working together has a galvanising effect when things get tough or momentum/inertia go against you. I’m very grateful for the people I work closely with.
Lots of this movement is broadly related to moving departmentally towards to a distributed data mesh and all that this brings. For people interested in reading more on this. This is a great long form read.
