On 7 March last, the Market Research Society (MRS) held a conference titled ‘Can Big Data replace the Census?’. It was an interesting topic that sparked my interest. The answer to the main question, however, is not as straightforward as you might expect.
I headed off to the seminar with an open mind. Even so, I couldn’t help thinking that despite the wealth of data available, it would be quite a challenge — if not impossible — to quantify the UK population in another way than through the traditional census. The speakers at the event echoed this impression. Many of them work primarily with other data sources while simultaneously relying on the Office for National Statistics (ONS) and its census data as the benchmark to calibrate against.
Rationalising and streamlining
Rather than being an attack on the census, the discussion revolved around how the traditional approach to quantifying a country’s population can be rationalised and streamlined. Or, in other words, how other data sources can supplement the information collected through the traditional ten-year survey.
The road to 2021
In the run-up to the 2021 census, the ONS is indeed identifying other data sources that will assist them in more accurately counting UK citizens and households. This will give them the opportunity to assess how well these data match what they collect in the survey.
The goal isn’t just to replace questions that appear on the census, but also to help ensure that the census is run as efficiently as possible. For example, investigating data from ‘smart meters’ and the Land Registry paints a clearer picture of occupancy levels in certain areas — and helps avoid the issue of chasing after responses from uninhabited properties.
It’s great to see how the UK government is increasingly harnessing the power of existing datasets to perform the census. This is sure to encourage organisations to make more data available. But with the existence of so many potential data sources, ensuring that data is clean and clearly defined will be the key to its usefulness. Data analysis needs to be built on solid foundations. Hopefully, the utilisation of additional data sources in such a high-profile publication as the census will lead to greater transparency and disambiguation — as well as to increased access to high-quality data resources that will further enhance what we at boobook can do.