Interview: ONS data science chief Tom Smith on government data capability

The managing director of the Office of National Statistics’ Data Science Campus, Tom Smith, describes the organisation’s data capability building activities for and across government

This article can also be found in the Premium Editorial Download: Computer Weekly: The real-life consequences of ransomware attacks

The Office for National Statistics (ONS) now has a higher profile than ever before because of its data dissemination role during the Covid-19 pandemic.

Its Data Science Campus, more particularly, is playing an increasing role in ameliorating data capability in government. One element in the government’s national data strategy – which has yet to fully emerge – was the goal of training 500 civil servants in data science tools and techniques during 2021. The ONS beat that target recently, having trained almost 700 analysts.

Heading the Data Science Campus is its managing director, Tom Smith, who describes himself as a “lifelong data addict with degrees in physics, artificial intelligence and computational neuroscience”.

Smith spoke to Computer Weekly when the campus was launched in 2017, and recently provided an overview of its strategy and outlook in an interview that complements an interview with Frankie Kay, interim director general for data capability at the ONS, last year.

Smith lauds the speech that Cabinet Office minister Michael Gove gave at Ditchley Park in July 2020, which staked out a high-level political and economic credo that took off, approvingly, from a quote – unusual for a Conservative politician – written in the 1930s by Antonio Gramsci, a founder of the Italian Communist Party. The speech was about what Gove sees as a need to revolutionise Whitehall, with data analytics as a crucible of that process.

The Gove quote from that speech Smith refers to is this: “We need to ensure more policy-makers and decision-makers feel comfortable discussing the Monte Carlo method or Bayesian statistics, more of those in government are equipped to read a balance sheet and discuss what constitutes an appropriate return on investment, more are conversant with the commercial practices of those from whom we procure services and can negotiate the right contracts and enforce them appropriately.”

Smith comments: “I think that Gove speech nails it, talking about the need for civil service, public sector skills to really strengthen around the areas of data science. It is a great illustration of how important data science skills and statistics skills are, and are seen to be, not just by data geeks like us, and me, but right across government, and at very senior levels. That’s the starting point.”

Building data skills

In relation to training and formation in data science skills, the ONS Data Science Campus, which is based in Newport, south Wales, is active in taught courses, apprenticeships and data capability building in government departments and across government. It offers a range of training, from a full MSc in data analysis for government alongside graduate and apprenticeship programmes, to workshops and development schemes.

Smith confirms that more than 50 civil servants have taken the Masters in data analysis, run in partnership with University College London and the universities of Southampton, Oxford Brookes, Cardiff and Glasgow, and that another 300 or so have taken modules from it. The apprentices include school leavers, but also career changers, of which there are around 45, he says.

“Gove nails it, talking about the need for civil service, public sector skills to really strengthen around the areas of data science. It is a great illustration of how important data science skills and statistics skills are, and are seen to be right across government and at very senior levels. That’s the starting point”

Tom Smith, ONS Data Science Campus

“But I want to get us beyond numbers and thinking about, say, targets of 500 analysts,” he adds. “I want to be thinking much more about how we support departments’ and agencies’ capabilities.”

The campus has run workshops with around 10 departments involving 250 or so civil servants.

It also runs a “data accelerator programme” together with the Government Digital Service (GDS) and the Government Office for Science (GO-Science), under the auspices of the Government Data Science Partnership (GDSP). This is a peer mentoring programme that pairs civil servants across departments to give them the opportunity to develop their data science skills. A typical recent cohort attracted some 90 applicants from across 25 departments and agencies, says Smith.

At a more senior level still, the Data Science Campus plans to support data masterclasses from this year. “This started out at Number 10 [Downing Street], and is a push for the very senior leadership across government to have a world-leading understanding of what’s going on in the data space – what the opportunities are and what they as leaders need to know about in order to make decisions, develop teams, be intelligent customers, that kind of thing.”

Number 10, working with the campus, has piloted this, with the campus now taking the lead to scale it up.

It is, he says, not about technical skills. “It’s about leaders understanding, identifying and having the kind of backup, that goes back to the Gove quote, to be able to procure, support and grow the [data] capability of their departments and their agencies.”

Value of data

Asked whether there is a danger that momentum could dissipate after the pandemic in respect of public sector data work, Smith comments: “I think that concern is a good illustration of why this [work] is important, including analysing scenarios of how organisations go back to work – office or not? – that sort of thing. How do we bottle what was good, rather than returning to the old ways of working by default?”

“Senior politicians and officials talking in great detail about data is a powerful example of how important this is”
Tom Smith, ONS

Of value, he says, has been the ability to share data and publish information very rapidly. “Having very senior politicians and officials talking in great detail about data, and about the analysis of it, is a powerful example of how important this is, and of why you need to invest time and energy into building your capability in its broadest sense, to ensure that it’s not just an emergency response.”

Data scientists at the ONS campus are also working on a plethora of analysis for the government. Smith highlights work to identify how the economy and society are emerging from the crisis, “using global shipping GPS data, traffic camera data, mobility data, satellite data and payments data to produce faster estimates of what’s going on, and that’s being published weekly by the ONS”.

The satellite data work includes keeping track of depleting water resources for the United Nations’ Sustainable Development Goals and working with the Foreign and Commonwealth Development Office to help local governments track livestock in South Sudan.

This year’s census will also be a key ingredient and a key source for data analysis and modelling. “As that data becomes available, [it will provide a] huge opening and window to look at a rich set of things, in the round, including when linked and analysed with other data sources,” says Smith. “For example, what do we know about inclusion, equality, inequality, diversity? The work on ethnicity and Covid that we published recently is the sort of analysis the census underpins.”

Read more about the Office for National Statistics

Read more on Artificial intelligence, automation and robotics

Data Center
Data Management