The deputy national statistician has said ‘three Ls’ – linking, localised and longitudinal – are going to be crucial features in government’s use of data over coming years.
Alison Pritchard, who is also director general for data capability group in the Office for National Statistics (ONS), also emphasised the importance of recognising critical national data assets.
She was speaking at the government DataConnect event, addressing the priorities for maintaining the momentum in obtaining value from data.
Pritchard highlighted the three Ls as key elements in the technical architecture for progress, saying there is now scope for a more ambitious approach to linking data.
“Our legislation is built around the concept of two parties sharing data and working with that, but we’re broadly finding there is a growing set of integrated data assets – such as public health data, something we’re growing around net zero and climate, the longitudinal outcomes data asset – we’re finding the power of that is exponentially larger,” she said.
Semantic mapping
This will increase the importance of the concepts of integrated data assets and semantic mapping, the latter making it possible to look at the data components of major issues such as climate change, levelling up and public health.
Efforts to combine and further develop the concepts are being led by the Integrated Data Service, which was set up by ONS last year and for which Pritchard is the senior responsible owner.
She said the importance of localised data reflects the need for a more granular understanding of several major issues, often going down to a ward level, and that can be highly important for local government.
This can also be aligned with personalised presentations of data, with an example being the ONS personalised inflation calculator.
For longitudinal data – collected from a respondent over a period of time – the major benefit is in making it possible to better understand the impact of a policy on an individual over the medium to long term.
Understanding outcomes
“Part of the ability for us to be able to evaluate policy proposals means we have to understand the outcomes that have occurred before with similar or differing policy information,” Pritchard said.
“It also allows us to do digital twinning work which will be fascinating part of data science where we look at different policy options and model those based on longitudinal data.”
She added that critical national data assets are becoming as curial as critical national infrastructure, and that when the two are integrated they can be very powerful. This will rely on a unified and effective data architecture.
“Imagine a world where we have object oriented data assets allowing us to share the entities and attributes across boundaries would be quite fascinating,” she said.
“We’re some way from there, but everything we are doing is gradually taking us there. As we improve our metadata, for instance, it opens up the opportunity to do that kind of thinking.
“But let’s not forget that underpinning foundational work has not been completed and we overlook that at our peril.”
Pritchard also predicted an increase in the amount of public facing data available for issues such as climate change and inflation.
