Weeknotes S1E2: A week of realising it’s all about the people (again)

I’m going to mix it up a bit and try and group my weeknotes by topic rather than by day. That way you can only read the interesting sections and skip the boring ones. How do you know which sections are interesting? Well, obviously you read all the sections first, and then only read the good ones.

Defra: Making Data Findable

Defra’s Making Data Findable project is all about providing Defra group with a world class data discovery tool that reaches across all of Defra’s agencies and bodies. This week we had outputs from the last round of user research summarised by Paul, highlighting his main findings and the problems Defra staff are having with data. It’s a really interesting set of results and goes way beyond just Making Data Findable, and has implications for how we implement Defra’s data strategy.

In particular for me it highlighted that users don’t just want to find good data, they want to find an expert who understands the data that they can have a conversation with. This idea of “surfacing the experts” is something I think has a lot of implications — firstly, the UX needs to attach greater prominence to contact information, highlighting who the key people involved with a dataset is. Secondly, there needs to more effort applied to getting contact information up-to-date in metadata and keeping it up-to-date — staff churn of data owners was cited as a serious problem because users don’t know who owns the data anymore. This is a process thing but it also means that however we use Making Data Findable to encourage owners to improve metadata quality, people associated with a data set need to be a priority.

It also suggests we might want to investigate other ways to help users “find data experts” — not just search but also more passive “browsing” of blog-style content that links domain experts with the datasets they recommend. It’s not the first time people have thought of this but now there’s some motivation to back it up. It’s a more direct encoding of the network of trust wherein people trust datasets because experts recommend them. This sort of thing is very common in scientific domains, which I was learning about from Hannah at the Grant Museum last week, and it would be interesting to explore where we could enable it here.

We also had a Team Excellence session run by our new delivery manager Andrew G. As a team we decided we’d like to be more like NASA, so we adopted them as our model team. I learnt that all of my team mates are considerably more interesting than I am! They’ve been movie stars, musicians, athletes and black-belts, and the session really helped me understand better them as people and what’s important to them at work. I also realised that when I’m at my computer I’m a terrible listener and I pay more attention to my screen that the person talking, so I’m going to improve that. If you catch me doing it, tell me off!

Defra: Authoritative data

As research has continued into Defra’s reference data sources, we’ve realised the issues we’re encountering are much broader than we originally imagined. We’ve also realised that calling something reference data is somewhat a matter of perspective, and not all the people we’ve done research with have connected with the term. So for now we’re using the words “shared authoritative data” to talk about what we’re looking at.

This week we spoke to some Rural Payments Agency (RPA) colleagues about their issues sourcing, processing and sharing geospatial data in the form of “data products”. We saw a real spectrum of datasets in terms of incentives for quality: land use data is some of the highest quality data around because there’s strong economic incentives to have robust quality processes, whereas some other datasets don’t have the same sort of resource and therefore aren’t kept up-to-date, which has an impact on downstream users referring to it.

When we make Discovery recommendations we’ll need to make sure that there’s a way for the business to show that a custodian is empowered to maintain the data they own — without that, they can’t be a custodian. We’ll need to think hard about whether that means the responsibilities of data owners needs to be split, and if there’s anywhere else in the organisation that’s already operating like this.

Defra: Data Target Operating Model

This week I got to contribute to technical part of Data Target Operating Model (TOM) which will be a “go-to guide” for what good technology and data practice looks like for our data services.

Obviously this draws extensively on Defra’s existing architecture practices, but also calls out important elements from the Data Strategy and Technology Code of Practice, like the use of open standards for data and the need for explicit data architecture input.

It’s only a first draft but I’m glad that when people are making new services in the future we’ll be asking the right questions for them to realise data is important and help them get it right.

Conclusion

Outside of providing consulting services to Defra around data transformation, I’ve also had lots of chats about data literacy and what Government should be doing about data infrastructure, motivated by some great discussions at last weekend’s UKGovCamp (which we sponsored!).

As with everything I’ve been doing this week, the conclusions of the discussions are consistent about one thing — it’s not about the tech, and it’s all about the people.