The numbers game: data and covid

By Gavin Freeguard, consultant on research, policy and advocacy work on data, information, media, digital, government and data visualisation.

--

In March 2020, the ‘flattening the curve’ graphic was everywhere.

It showed an early, high peak of covid-19 if there was no intervention; and a later, lower summit if we successfully socially distanced. Dotted lines showing our ‘health system capacity’ and other helpful labels appeared.

Copies, critiques, praise and pastiche followed. As calm gave way to a storm, the curve gave way to the daily dance of charts macabre, the logarithmic waves refusing to break as death tolls climbed. The UK government’s daily covid-19 press conference brought a rhythm to the news cycle and with it, new charts on transport use and testing.

The pandemic has been dominated by data. Charts and numbers have been at the forefront of public debate like never before. It is a reminder that data matters. In and around government, the focus is (understandably) so often on data policy, standards and ‘fixing the plumbing’—so much so that we forget the importance of communicating and visualising data.

What lessons should we take from the data developments and debates of the pandemic?

There’s years of work behind that overnight success

The crisis accelerated existing trends and built on firm foundations. The ubiquitous Financial Times charts were both of the moment and a result of years of building teams of data journalists and finessing charts and graphics. Other publications, civil society organisations, thinktanks and some parts of government have taken data visualisation more seriously over the last decade. And that’s been evident.

Data visualisation is a ‘functional art’

That said, data visualisation still leaves much to be desired. Too often, decision makers are tempted by flashy graphics at the expense of functionality, forgetting what it is they’re trying to communicate.

Take the broadcast one Sunday in May 2020 where the Prime Minister spoke of easing the lockdown. The public’s assumption—quite understandably—was that the series of dials, scales and maps that accompanied his speech showed the equation used to work out the country’s (newly lowered) covid-19 alert level. They did not. They were as likely to confuse as they were to clarify. And the less said about the UK Government’s notorious October press conference and its unstructured, ugly and unclear presentation of data, the better.

On the other hand, there are still enough badly distorted default Excel charts in significant government reports and press conference slides to suggest their compilers are dismissive of the art. There is nothing to entice the reader. In being dull and merely functional, they fail to fulfil their function: of communicating something to someone.

You wouldn’t throw letters on a page and hope they ended up as clear, concise communication; the same is true of throwing numbers into charts and hoping they make visual sense.

Make things open, it makes things better

Many of the most memorable charts and most acute analysis came from those outside government. However skilled government analysts are, there will be others on the outside who can bring fresh perspectives and different skills—and should be able to check government’s working for themselves and the wider public. A decade or so of open government data meant a lot of relevant information became available, but greater openness globally could have helped fight covid-19.

Many complained as the UK government changed how it published some covid-19 data as the crisis unfolded—any open data is better than none but consistent, good quality data is vital for use and understanding.

One trend that will hopefully continue is the willingness of the UK Statistics Authority to intervene. As well as criticising muddled and misleading government numbers, it has also sought to force the publication of underlying data cited by politicians.

Improve our understanding and acceptance of the limitations of data

Theoretically, there is a definitive number of people killed by covid-19. Practically, it is a number beyond human counting, if not conception.

There were at least 3 prominent ways of counting deaths in the UK:

1. Daily deaths (from Public Health England and relevant bodies in the devolved nations).

2. Weekly deaths (from the Office for National Statistics, using death certificates).

3. Excess deaths (those above what one would normally expect).

The strength of the daily numbers was their closer-to-real-time sense of what was happening, the weakness was their lower accuracy (not least when they only included deaths in NHS settings, and later included deaths from those that had contracted covid-19 but recovered before dying). The weekly and excess deaths may have got closer to the unknowable true figure, but they took longer.

It would be easy to blame the government for a lot of the confusion about these different numbers. And there is much to blame them for in the way the numbers and their limitations were (and were not) communicated. But some of this is also on us, the public, because in a time of great uncertainty, we sought a certainty in numbers that was simply impossible. This is a close cousin of the tech solutionism of seeking solace in ‘easy’ technological fixes, evident in the fanfare around the NHSX covid-19 app and the algorithm used to distribute A Level results.

Life, numbers and data are more complicated than that.

Avoid ‘number theatre’

Far from instilling greater numeracy in the population and promoting considered debate about the merits and limits of data, the government’s daily press briefings instead treated the public to ‘number theatre’, which was ‘not trustworthy communication of statistics’, according to statistician Professor Sir David Spiegelhalter. In the place of genuine information came spurious precision on deaths and tests. Consequently, the public drowned in a deluge of statistics. More information became an impediment to understanding, not an instrument of it. As with designing effective charts, we all need to ask ourselves what is it that we actually want to communicate, and what is the most effective way of doing so? The pandemic has put the use and communication of numbers in public life under the microscope. But in magnifying existing problems, it offers us all the chance of a cure.

public digitalThe public digital logo

Head Office

Clerks Court
18-20 Farringdon Lane
London, UK
EC1R 3AU

Our positions

Our values expressed in action and outcomes.

Read them here

Newsletter

A monthly scan about digital transformation and internet-era ways of working around the world.