When I started the hub my vision — although I doubt I would ever have referred to it as such — was to develop
- A suite of applications that — worked individually for specific user groups to address now needs — worked together soon to address strategic needs — worked with other products outside our team to eventually join things up more for data consumers.
- A suite of analytics pipelines that — could be embedded in the applications — addressed the most commonly asked high level questions — could be tailored and adapted for specific low level problems.
- A-Team (you see what I did there) with an R&D culture and capability.
All this with a principle of reusable micro-services, common design patterns (for 1 and 2) and most importantly (for 3) pragmatism and respect for where the organisation was, is and is going.
Its starting to come together
This week has started the conversation about strategic needs. However there is still got a long long way to go.
I said that I anticipated this being the final episode of this season — mainly because I hoped to be on leave for a few weeks after this . As it happens I’m around for half of next week so I’ll put in a bonus #halfweeknotes next week which will include presenting at the Government Data Science Community of Interest which I’m looking forward to.
Week in brief
I was down in London for a meeting with the Minister for Employment.
I travelled down with Becky (data product owner for Manhattan). When we arrived we caught up with Charlie to run through what we were each covering. Charlie then gave us a bit of an update about the Data and Analytics senior leadership team meeting. Some of this was linked to my thinking cap point in episode 6.
While Becky familiarised herself with the QlikSense version of Manhattan on my SurfacePro — some things are ever so slightly different from her machine and the train wifi didn’t allow us to test while travelling — I took the opportunity to have a quick catch up with Iu about a few things. He reminded me — again — that he was going on leave for three weeks. He also prompted me about the joint good-cop bad-cop blogpost on data visualisation we’ve spoken about.
We popped upstairs and met Pauline C (Deputy Director, Labour Market) who filled us in on a few things. I’ve seen Pauline talk at a few things before and she’s brilliant. I’ve only actually met her once, well over a year ago but she remembered which was nice.
Into the meeting. We gave an overview of the Churchill, Greyhound and Manhattan applications. Their individual purposes and also demonstrating how they all work together. A couple of minor issues with wires and wifi but everything was well received and some good conversation and outcomes relating to the use of the applications for strategy. A quick debrief afterwards with Pauline and Charlie. Pauline said something which was great to hear about the applications
they just make it so easy to answer questions
I chatted with Iu about modelling with encrypted data — I’ve been thinking about a name for this piece of work but haven’t quite made my mind up yet. Iu had spoken positively with one of his contacts and made some additional ones so we’re gathering momentum on what is essentially a — potentially very valuable — R&D side-project at the minute.
It was a bit of a flying visit. On the train back Becky and I chatted through a few ideas and I cleared some emails. It was Sree’s (data scientist) last day in the team which I was sorry to be missing, we all wish her well in her new role — which will be in another of Charlie’s teams.
I was back in time to meet my wife and kids at a new climbing place.
When I arrived in the office, Gayll (data scientist who oversees Greyhound and Manhattan) wanted a chat about a meeting she’d been invited to on Friday— Universal Credit Digital Transformation senior leadership team meeting — chaired by Deborah B.
The invite stemmed from a series of conversations and interactions which I mentioned in episode 2 — Gayll had been leading things since then and wanted a quick chat about the scope/expectations outlined in the agenda since it covered a few things outside of our area of responsibility and asked if I’d join the meeting. We set up a team chat for Thursday and Gayll was going to gather some info from Iu and Luke.
A quick change of buildings and into the team show & tell and retrospective. A different set up to normal as Stephen had organised a demo of a data blending and advanced data analytics tool. I then had to leave to call Philippa just at the point it turned into the regular meeting.
I had a really great chat with Philippa who I met in episode 4 where I said I wanted to follow up and learn from her experiences as a (non data scientist) product owner/manager in/of a data science team. You may remember the subject reared its head again in episode 5 after I read Zoe asking what does a product owner own anyway and Louise talking about Directly Responsible Individuals in her #weeknotes.
Philippa and I ended up talking a little more widely about what we each do and shared our thoughts and views on a few subjects which included.
- Differences between — products that need/use data — data as a product — data products. More to discuss on the latter.
- Differences between product owners and product managers and how, confusingly, these can alternate between departments.
- The importance of behaviours and digital/agile mindset — whatever you define that as — plus multidisciplinary teams when it comes to data products — whatever they are defined as.
I love chats like that. Philippa subsequently sent me a blogpost link which contained the following which I liked.
these days, it’s not enough to just toy around with data, creating models that end up in a powerpoint presentation but they need to be part of a product that is shipped! That’s why good data scientists need to be good product managers as well. Good data scientists need to understand how their work end up in a product and how it will be consumed.
I’m still thinking about this subject but the chat made me even more clear that — to quote Philippa
it’s behaviours not career histories that make good data product managers
I hope to chat to Zoe and Louise for their views — plus Zoe and I need to chat about her masterclass idea! Theres more to come on this I feel.
I caught up with Stuart (data product owner for Churchill). He was delivering a workshop with pensions policy folk in London the following day. I was supposed to be joining him but Charlie needed me to cover some interviews in Sheffield instead. Stuart wanted to check he wasn’t missing anything. He’d had a crash course in Greyhound from Daniel (data scientist) and we covered some parts of the wider narrative.
Alaine (delivery lead), Dave D (business analyst) joined us for a quick chat about working with users.
I joined a Lync session which Becky had lined up which covered a potentially very useful data source for us. This could be useful in its own right but also could be really useful in breaking down some of the barriers between my teams products and Kit’s — part of a wider vision for a common data currency between operations and strategy/policy.
I have to apologise for this next gif — I felt like I’d committed to a bit of a retro tv/film theme with Hannibal and Axel and then I wrote breaking down barriers and Kit and then this happened.
I had a few conversations covering the Friday meeting in Harrogate and — unrelated — postcode shape files. In between I popped downstairs to buy a cake from Ben, George and Ash (developers) who are taking part in Tough Mudder and raising money for charity.
In my notes for the day I wrote
feeling a bit swamped at the minute
Interviewing in Sheffield with Billy (head of the Sheffield hub).
Before that discussions with Alaine — on the phone on the train down — and then Max (program manager for Data Science) — when I arrived in Sheffield — about dependencies between our work and data architecture to feed into a Data and Analytics senior leadership team planning meeting.
Back-to-back interviews all day for data scientist posts in Billy’s team. The final candidate didn’t turn up so I got away a little earlier. I phoned Stuart on the walk to the station to see how it had gone with pensions policy people. It turned out our trains were arriving back in Newcastle at the same time so we met up for a couple of beers.
My son was really quite poorly through the night. For various summer holiday logistics reasons this meant I needed to be around in the morning until my wife was back to look after him so I didn’t get into the office until the afternoon.
I met with Gayll, Becky and Daniel about the Harrogate meeting. We agreed the content, who does what, timings and the outcomes we’d hope for — Gayll had collated some views from elsewhere for us to incorporate.
I was double booked next. I joined the IT health check update for Churchill private beta first. Outcome is, no major issues, still on track so should be making the application available externally to Kate and a few others very soon.
We talked about the icon for the application. This is a draft of the current favourite which we think encompasses most of what we want to portray.
The update finished early so I was able to join the second half of the meeting about Mercury — another R&D side-project — which Stephen and Daniel have been working on with Dawn S and Ian J (Data and Analytics Private Office). Mercury is looking to enhance the information they use and have access to. Some practical outcomes which Daniel and Ian will work together on.
Harrogate for the Universal Credit Digital Transformation senior leadership team meeting.
On the train, Daniel, Gayll and I talked about planning and I joined a call about resourcing with Charlie.
Daniel familiarised himself with the QlikSense on my SurfacePro before we went into the meeting. We went in with the usual array of wires, connectors and cables but still needed Heather to help us out!
Deborah and her team were very welcoming and asked really good questions. We presented the applications — a similar presentation as given to the Minister — and touched upon the common currency I mentioned earlier. A really good meeting with positive outcomes and actions.
What is hard?
Being in the middle of a data and digital culture change venn diagram. The data product owner questions I’ve raised in these posts are a very tiny example (see also data visualisation designer) but more widely facilitating perspective shift is hard work — especially when it’s about data.
What is fun?
Working as part of a multidisciplinary team with a fantastic attitude to problem solving, challenge and change. We always find a way. As Stephen says
when things go well it’s fun, when they don’t it’s funny.
What makes me proud?
The reactions to the products. As Pauline said
they just make it so easy to answer questions.
That was my week.