Although the start of this year has been a bit beset by colds and flu (mine, and toddler’s - which has led to a lot more sitting on the sofa watching Frozen II in the last 10 days than I would usually have!), it seems like a lot has also been going on.
Building the research team
We had the first of our weekly Research Team meetings yesterday, trying out a format based around four questions:
- How are we?
- What are we working on?
- What are we learning?
- What are our questions for the week ahead?
We gave about 10 minutes to each question, and took time before the questions for a few minutes of quiet reflection, and putting the points we want to raise into the note-taking doc ahead of a round of inputs, and discussion.
For each of the questions we could reply at any level of granularity. For example, a learning point could be something focussed and practical (e.g. how to tweak our online survey design to be more accessible), or broad and open (e.g. a new conclusion from work, or discovery of a new concept).
I think it worked reasonably well to mix both project management and support, making connections between our strands of work, and allowing us to sense some of the wider issues, trends or resources our research work could address or draw on in the year ahead. We’ll keep iterating on the meeting model.
Exploring Insight Infrastructures
We’ve been doing a lot of ground-work for our project with Joseph Rowntree Foundation on Ecosystem Mapping and Engagement for their developing Insight Infrastructure project.
Archetypes, outliers and keystone species
As we’ve been building out a list of organisations involved in creating or using data and insight to address poverty in the UK, we’ve been grappling with how to scope the mapping work. Clearly we can’t map (even a fraction of) all the organisations and communities in the UK that address poverty, social and economic inequality, in order to explore how they use, or need, data, information and insight for their work. However, we also need to avoid being scattershot in our approach, and to make sure we’re learning from interesting examples.
My working approach to this, which I’ll be testing more next week, is to first think about mapping each organisation to an archetype.
For example, Surrey-i, which I was looking at this week (thanks to signpost from a member of our Discord community - do drop in), is one instance of a Local Information System (LIS): as far as I can tell most County or Unitary Authorities have one.
We ideally need to have a fairly exhaustive understanding of the ‘archetypical’ actors within the ecosystem we’re looking at: but we don’t need to survey every single instance.
At the same time, we do want to identify ‘outlier’ examples of some archetypes that are going to tell us something interesting about the challenges, or opportunities, of creating or using insight in a particular setting or context - or that represent particularly interesting stakeholders to engage in future stages of the project. I’m drawing here somewhat on past conversations with Andi Pawalke around the importance of tracking ‘positive deviance’.
And as we sample a few instances of each archetype, I’m discovering we often run up against some common organisations they relate to. For example, Local Information System’s often draw on analysis produced by the Local Insight platform from OCSI. This points to the presence of a ‘keystone species’ in the ecosystem: organisations playing a central role to sustaining critical activity of many others.
Alongside responding to a brief from JRF, we also have a series of questions we’re hoping our work on the project will help us to explore. After we met last week for a cross-team brainstorm, I had a go at conceptualising these questions in more detail, thinking in particular about how they relate to the components of good data governance.
1. [Collective] How do we effectively introduce collective concerns into the design of data systems?
We are seeking to identify questions and perspectives that might be introduced during the process of developing data and information systems in order to surface collective impact issues and approaches.
For example, our survey for the Insight Infrastructure project invites respondents to identify particular data resources they use or create, and asks: “Are there particular groups that would be impacted by wider use of these information, data and analysis resources? If so, how?”
The hypothesis we are seeking to test here is whether asking ‘experts’ to consider groups impacted by collection or use of data will reveal particular communities that should be considered and/or involved in the governance of data.
2. [Participatory & Democratic] Where and when should communities be directly involved in the design and development of data systems?
In any evolving and literature project there will be questions about how the direct involvement of affected communities in the design and development process should happen:
- At an early stage, it may not be clear which communities are affected by the proposed infrastructure, and so unclear who should be in the room;
- Dialogue when plans are early stage may be less likely to be powerful, as there will be many constraints/existing agendas that are implicit, making it difficult for public stakeholders to provide informed and targeted inputs;
- If communities are involved too late, then decisions may have already been made that constrain the kinds of things that community voice can affect.
We’ll be reflecting on the different ‘sites of engagement’ and ‘moments of engagement’ that projects like the JRF Insight Infrastructure might afford.
3. [Deliberative and Powerful] What are the appropriate models for governance of a shared insight infrastructure?
To some extent, governance follows form, follows function: and there are many possible institutional and governance issues and models for an Insight Infrastructure. We’ll be thinking early on about the kinds of governance models that might be relevant, and using interviews and workshops to understand the strengths and weaknesses of each.
[Political] 4. How can an engagement and communication strategy support good data governance?
The way a project is described to stakeholders and contributors, and the way they are brought onboard will have implications for how data and insight is governed. If communities are to have a powerful voice in shaping the inputs or outputs of an insight infrastructure, how does communication manage expectations amongst different stakeholders to make this a source of strength, rather than tension?
Of infrastructures and ecosystems
Reading Jeni’s post on ‘What is insight infrastructure’ got me thinking about a dimly remembered paper I wrote whilst at the Web Science DTC in Southampton. I tracked down a copy, and whilst lots seems a bit dated/superseded, I thought the following was still useful (albeit explicitly about data infrastructure rather than insight infrastructure):
The deployment of twin concepts of infrastructure and ecosystem to describe technical and sociotechnical systems has widespread precedent. This pairing allows a distinction to be made between infrastructure as the basic physical and organisational structures and facilities needed for the operation of a society or enterprise (often centralised, standardised and managed by some small set of agents), and the emergent, autonomous and self-organising components of an ecosystem, linked together in local and global feedback loops and developing according to local specialisations and adaptation rather than top-down design.
Grounding digital rights
I had a fascinating conversation with Professor Tom Stoneham from the University of York who has been developing a new MA Course on Applied Ethics and Governance of Data Privacy, to be taught for the first time in September this year. As an academic philosopher, Tom was posing not only practical questions of how to equip ‘data privacy professionals’ to ask and operationalise deeper ethical questions about how data is managed in government or industry, but also exploring how claims around digital rights might be philosophically grounded.
In particular, Tom noted the gap between universal (but often general / high level) human rights, and contingent (and often very specific) legal rights, that can ultimately be changed by legislatures at will. The challenge then is how to ground ‘rights’ claims (such as a right to encrypted communication) that may not be easy to articulate at the level of human rights, but that should not be solely at the grant of national legislatures. I’m not sure I can reconstruct the full sketch Tom gave of how to address this challenge, but as I understood, it invoked consideration of the basis for a (particular) state’s legitimacy, and rights acting as a buttress to actions that would otherwise undermine that foundation. This creates the space for a broadly pluralist account of digital (and data) rights, in which different rights might be articulated by different countries and political communities.
While my undergraduate training leaves me wanting to dig much deeper into this, and follow a number of the philosophical alleys this opens up - at a more practical level for Connected by Data, this connects with conversations we’ve been having about the role of ‘legitimacy’ as a concept, and the role of community voice in giving or withholding legitimacy from particular data governance choices. I’m hoping we’ll be able to have a Connected Conversation session on this later in the year, and bring Tom into those discussions.
Other things from this week
I contributed to a proposal led by Aapti, and along with GPSDD to provide a policy paper into India’s T20 process that, if selected, will give us an opportunity to feed ideas of collective and participatory data governance into upcoming G20 meetings.
We had a check-in on planning for the Gloucestershire Data Day - now confirmed for 26th April 2023 at the Churchdown Community Center. Plans coming together and more info (+ call for other sponsors / contributors) should be out in early Feb.
I’ve been doing some updated work to explore how we have a simple CRM for Connected by Data that allows us to easily implement our data protection policy. I’ve still not found an off-the-shelf option that elegantly handles tracking the legal basis for information processing, supports easy deletion, and is designed for data minimisation rather than invasive data collection. Pointers welcome, but feeling like I might have to return to my ‘Plan b’ of Airtable + custom data management scripts.
I added Mastodon links to the website template (and Jeni sorted the formatting! Thanks!)
Next week I’m looking forward to being in York on Monday for a workshop organised by Wellcome and Involve that is asking “How can data policy be informed by sustained and sustainable public engagement?”, and I’ll be digging into our case study collection to put together a lightning talk for that.