This has been a week of many threads, some starting to weave together, others just emerging as threads to pull at and explore more. I’ve tried to capture some in the bigger block of reflection below, in addition to which, the last few weeks have involved:
- Hiring! Thank you to everyone who applied for our Research Associate role. I had a tough time getting a list of 35 applications down to a shortlist of 10, and preparing for a first round of interviews next week.
- New funding. More news on that next week, but it means I’ll be in the happy position to hopefully hire two rather than one research associates!
- Testing new tools. Trying out Research Equal’s collection tool, and putting together a quick collection of papers on Collective and Participatory data governance.
- Working through interview transcripts for our evaluation of Justice Data Matters, and using the OGP’s framework for Informed Participation to look at strengths and weaknesses of the polling and dialogue process used.
- Catching up with Kasia Chmielinski from the Data Nutrition Project on how the process of labelling datasets for machine learning use can incorporate voices of affected communities. I found the existing field-level classification of potential individual and collective harms that might exist in a dataset particularly interesting - and raising some interesting design challenges for how to move that from being expert defined, to more community identified.
- Our in person team meeting in Reading last week, where we dug into the policy positions we’re advocating for, and those we more or less reject. In the next few weeks we’re going to start sharing some of the critical questions we expect to get asked about our position on the Connected by Data Discord, and hoping to shape some discussion on how they can be answered.
Relocating power: structuring data ecosystems
What does it mean to secure more equitable and effective governance of data in particular fields? Can we add collective and participatory governance processes onto existing data architectures? Or do we need to rethink data architectures and institutions at a deeper level?
Over the last week I found myself talking to Mariana Broitman at Fairfood International about approaches to give producers a greater stake in governing supply chain data. I talked with Shannon Dosemagen of the Open Environmental Data Project, about where local grassroots environmental monitoring meets EPA data collection processes. We caught up with Natalie Byrom as we started to write up findings from our evaluation of the Justice Data Matters dialogue on courts data. We’ve been chatting with Diarmaid at Just Treatment about the challenge of translating public views on health service data use into meaningful governance of data. And today I joined a fascinating workshop in London organised by CIFAR that explored how womens’ agricultural cooperatives in India might leverage data to improve their access to markets, and that included examples of co-producing data infrastructures and governance from Emma Beck at Equal Care Coop, Krishnan Pallassana at Digital Green, and Astha Kapoor and Shefali Girish of Aapti Institute, amongst others.
In all these conversations, I’ve found I’m coming back to two themes:
Firstly, the need for points of collective data control that are closer to affected communities. What does that mean in practice? Well, so often, data architectures work on a hub and spoke model. Data flows from individuals and communities direct to some central hub, with no local points of aggregation, intermediation or control.
Separated individuals have little power to gate-keep data collection, advocate for data structures that serve their needs, or set the rules that govern data sharing.
By contrast, when, for example, a group of farmers first pool data at the level of a farmers co-operative, or regional producers organisation; or when shrimp fishers come together in a given community to talk about the data they are collecting on environmental conditions, there is greater opportunity both to use and benefit from data locally, and to gain a more powerful collective voice in data governance.
Does this mean building new data institutions? Data trusts and coops? Not necessarily I think. The CIFAR workshop in particular explored the role of existing structures, like cooperatives - where data governance conversations can layer onto existing relationships and trust. And with plans for next years’ Gloucestershire Data Day we’re thinking about how just bringing people together to recognise that data collection requirements placed on them are not wholly given, and can be debated.
Second, the need for sector specific strategy to embed collective and participatory data governance. Agriculture, water quality monitoring, air quality monitoring, court records, arts and health evidence collection, public health data, patient records - data is always about something, and these ‘somethings’ all have their own contexts and cultures. As we’ve been thinking about the kinds of resources Connected by Data might provide to better support collective and participatory data governance, it’s been becoming clear that the need is less for ‘best practice’ templates or guides, but more for strategic thinking on how to build community voice sector by sector.
We’ve been putting together some ideas around ‘design lab’ events that could help in developing these sectoral strategies, and starting to plan a workshop with Legal Education Foundation as part of our work on Justice Data Matters to explore strategic next steps to build on learning about public demand for a say in how court data is governed.
Lots more to unpack on these themes (and some diagrams to draw!), but my train is pulling into Stroud station, so that will have to wait for future weeknotes.