The government’s recently published Transformation Strategy addressed the public sector’s need for an effective data strategy. We’ve highlighted how this has been lacking in recent months, with the required leadership seemingly remaining unfulfilled.
Data sits at the centre of the Government-as-a-Platform agenda, and underpins much of Whitehall’s transformation plans. However, data in government is messy, siloed and fiercely protected by departments.
There is a need to boost the data skills base in central government to help improve the current situation. The last government data chief, Paul Maltby, who has since moved on, told MPs over a year ago that he needed data scientists and he needed them quickly.
Part of the plan to broaden the talent pipeline for data skills is the creation of the new Office for National Statistics (ONS) Data Science Campus - which the Transformation Strategy describes as a “national hub for data science skills”.
The Campus has only just been announced and opens its doors next month, so it’s still very early days. However, Tom Smith, who has been appointed MD of the organisation, was speaking at the Think Digital Government conference in London last week, where he gave some insight into the organisation’s plans.
Smith’s discussion broadly focused on two priorities for the Campus: 1) improving the data skills pipeline across the public sector, and 2) thinking differently about the government’s approach to data and statistics, through the use of new approaches.
We are very much in the start-up initial phase. We launch next month, so we are not quite stealth, but essentially a lot of the work is being set up in its early days.
The ONS has a very simple task: to produce statistics for the public good, for the UK. That’s information that underpins decisions by government, underpins decisions by businesses, by the charity sector, underpins understanding of the general task. There’s a huge amount of energy and work, and a huge amount of challenge, as you’d expect.
What we do want to do is continually improve. Want to step up and step on the gas - one of those areas is to increasingly improve the way that the ONS looks at the datasets that are available and used for understanding the economy, understanding society, but also the methods and techniques that the ONS has used.
What’s this got to do with digital government? I see data science as a fundamental and essential part of the skill set to understand the digital world.
Smith said that data science sits at the intersection of three areas - maths/statistics, software development processes (the ability to use tools to cope with big datasets), and substantive expertise.
He added that if you bring the maths/stats together with the software development processes, this is where you ended up in machine learning territory. However, Smith said that you want to ally this with substantial domain expertise. He said:
At the intersection of those three, is what you would call data science. This is the exciting thing for the ONS, to push on, improve, and strengthen the skills around data science and data scientists in ONS and across government.
Smith said that whilst there are a growing number of people across government that would call themselves researchers, or mathematicians/statisticians, the ONS Data Science Campus wants to boost the pipeline of data scientists. Smith explained that the Campus is doing this via the creation of apprenticeships and providing Master of Science (MSc) degrees to colleagues higher up the value chain. He explained:
We are looking at increasing the pipeline. How can we increase the skill level? How can we increase the number of people coming through the system?
We’ve started early with this, we think this is unusual. We’ve started looking at apprenticeships. We looked at running the first data science apprenticeship in the UK, possibly the world. It’s a two year vocational programme. What will this bring in is people that are ready to go and work in government.
We had a huge amount of interest. We advertised this towards the end of last year and had about 140 applications. We took on 8 to start with, from a very interesting mix of backgrounds. We are shortly recruiting for another sample, so we expect to double that again.
Interestingly, they’re already starting to have an impact. They’re already working on projects.
Smith added that the ONS is looking to run its own MSc course in data science, for those with some experience/qualifications in the subject area already.
The second element to the Campus’ agenda is to experiment and try examine new ways of collecting and analysing data. Smith gave an example how Afghanistan hasn’t had a population census carried out in a long time, for fairly obvious reasons, and as a result it is very difficult to know how many people live in certain areas. This means that it is difficult to know how to best provide basic public services, resource allocation is tricky.
However, it recently ran a census based on the innovative use of satellite data and image recognition. Smith said:
We don’t need to do that here, but what that shows is that you can use a totally different type of data source, a totally different type of methodology, to produce an output that is hugely important to policy makers and business decision makers.
Smith explained that the Campus is working with users to make sure that the data they’re using and the data they require fits with how they’re making decisions on thee data they need. Some of the examples of its upcoming work include looking at tourism data at a local level (who is coming from where and to see what), working with Defra on industry classifications (what types of industry are based in which areas), and working with the health sector on looking at problems with data as it relates to Britain’s diet.
There’s a coherent story across them, which is about new data and new techniques. Trying to understand the economy and society in a better way.
We are experimenting. We are running short projects. Nothing more than 6 months. My view of government from the outside is that lots of stuff is going on that doesn’t see the light of day. Doesn’t have a defined end point.
My experience from the commercial world is that you need to be very clear on what you’re trying to do. What does success look like? And if we get there we stop and if we don’t we stop. We don’t let projects just roll on.
We are always looking to work in collaboration. We’ve signed a huge number of memorandums with universities, with businesses that may be data holders, and groups like the Alan Turing Institute.