Analytics official says Whitehall has to deal with skills and ethical issues to extract the full value from big data
Central government is still some way from extracting the benefits of big data, not least because of a lack of data science skills, according to the head of data and analytics at the Government Office for Science.
Chris Fleming emphasised that he was speaking in a personal rather than official capacity at a techUK event on the subject; but he was clear in identifying a group of obstacles in the way of departments making full use of the data at their disposal.
His comments have come as there is growing interest in private and public sectors in how to find insights from bringing together vast datasets from different sources. This is giving rise to a demand for data scientists who can analyse data that is often in unstructured formats.
Fleming said there are four main challenges, the first of which is that the cohort of traditional analysts and statisticians in government lack the IT skills to match those in maths and problem solving.
“The second challenge is about customer pull,” he said “There’s no point in building capability if there’s no market for it, and awareness of what these technologies can do has to be built.
“We’ve improved it at senior levels of government, but the challenge for us is in that vast swathe of middle management, in helping them become intelligent customers for data analytics. We have to find our data champions in departments and work with them to get real culture change.”
The third area is in what Fleming described as “the social contract on data use” around trust and reciprocity: government has to be clear about what people will get in return for providing their personal data.
Fourthly, decision-making in government is very complex and has to be based not just on data, but in a “sweet spot” between evidence, policy-making and delivery, and in another between security of supply, sustainability and affordability.
He also warned that, while the Civil Service is recruiting some people with high level skills through the Fast Stream programme, their capabilities may soon be blunted by the lack of appropriate IT tools. This could produce an “unvirtuous cycle” which makes it harder to build the capability.
Fleming said, however, that the government pushed the issue up the agenda with the creation of the Government Data Science Programme in January 2014, and that the relevant work is now under Mike Bracken’s wing in his remit as government chief data officer.
The programme includes work on a relevant skills framework, demonstrator projects, the creation of a community of interest for data scientists, an accelerator programme to lead analysts into data science, and work on an ethics framework for data scientists. Fleming’s team is also looking at how to establish best practice among departments that are using cloud based platforms for processing big data, and the Office of National Statistics is building big data laboratory.
“The plan for this year is to continue along these lines and trying to scale up and focus on real savings,” Fleming said.
Image from gov.uk under Open Government Licence 3.0