Sprint Notes – CMD Beta – Sprint 8
Such is the life of a service manager, I am writing this on a small train somewhere near Bath. The trees outside the windows are on the turn and the world has an unmistakable sense of autumnal vibes about it. Blimey! That means we are quite a long way into the Beta phase of the Customise My Data project and halfway through our (financial) year-long bid to continue to improve ONS’s website.
Let’s take a look at the last two weeks and maybe add a sprinkle of what that means for the bigger picture
First up – I got excited and dropped a rare GIF on Twitter cheerleading about the great work the team had just finished to ensure we have a working end-to-end development environment with real data flowing through it.
This is a good moment on the project, but comes with a point of reflection. We should have done this earlier. We had a working system about a month ago, but got distracted and didn’t make sure we deployed it to a working environment and then started iterating on things again. A note to Future Andy: make sure everything gets deployed to working production environments from Sprint 1 on future big projects. It will save time and effort (Future Andy is so wise)
In other news
The ONS Digital cake blog saw some major developments, but even the delivery of the first savoury item of the season couldn’t deflect away from the range of things we actually manage to build as well.
The team has been out and about speaking to folks. We ran some user research in Bristol, Southampton and London (and Benjy has been to seeing our civic tech brethren at the Parliamentary Digital Service) and have been processing a report that the wonderful folks at the Neath Digital Accessibility Centre (DAC) have pulled together for us. This volume of our user research is a time consuming thing, so once again, I am in awe of Al for managing to keep the show in the road. The feedback we gain from these sessions is the most important optic for how we iterate the development of the service..
Last sprint we spent time building some prototypes around geography, but in this one we have been looking at a couple of other important areas. Part of this has been the process for importing data into databases (something we will do via our bespoke CMS, referred to as Florence) and work on what a “landing page” needs to look like for external users of the new functionality. This is particularly important, as one of the main drivers for this project is to make our data easier to find, so making sure we have the right metadata available early in the journey is really useful in allowing users to understand if they are in the right place.
We have been continuing to build the collection of APIs that underpin the ability to breakdown datasets into smaller elements and have, for the first time this project, started to engage with some of the complexity inherent in sparsity within data. At the highest level, how do we guide our users through the complexity in data not being uniform? Sometimes, parts of a dataset could be at a lower level than others and also, how do start to build in the logic to deal with suppressing disclosive data. We will be talking about this an awful lot more in the future.
Alongside the CMD project, we have been continuing to work on a wide range of other features for the site. Crispin has been working his way through a really nasty bug in our charting tools and Benjy and (twitterless) Awen have been doing some great work on A/B testing. In the coming days you may be lucky enough to be in the 10% of users who will see some new routes through our product pages as part of our continued efforts to integrate all of our content into one site – currently some things are published on Visual.ons.gov.uk and we are starting to bring that back to the main site.
Finally, we are gearing up for some of the biggest changes to our production platform since the launch of the website 18 months ago. This won’t have any visual impact to our external users, but we are going through a range of security assurance processes to ensure that the refactoring we have done continues to be to the standards we require.