The Scottish Census Table-Builder

In which I continue to play around with data in the census and have a great time

The Scottish Census table builder utility allows us to create datasets from any combination of questions in the census, to find interesting correlations.

For example the question about general health and the question about Scots language ability - People who reported “Very Bad” general health are 50% more likely to speak Scots than people who reported “Very Good” general health.

That’s the attention-grabbing headline.

Ways to display proportions of Scots ability

The basic data output from the the census looks like this. If you like tables of numbers then its a party, if you’re not really into data, then its is pretty boring to look at

If we conceptualise the various language skills as being concentric groups we would be wrong, but it looks nice, easy to understand.

Although this maps out the 5.5 million population of Scotland as something like a Euler diagram, it is wrong because the various groups overlap in different ways. There are people who speak Scots who can’t read it and there are people who read it who can’t speak it.

The “Speak only” and “Read only” are roughly the same size, and each make up around 3.5% of the population, they are significant groups who have some skills in Scots but not the whole set of skills.

Here is a slightly different version, which uses proportionately sized grey blocks to accurately display the different proportions of the population with each Scots language skill.

Does it make sense? Even without the text saying what percentage each block is, you can still see that the “No skills” group is the largest, the “Speak, read and write” group is the next largest, and so on. Hopefully this is a little intuitive.

Scots skills in different age groups

The census provides data on Scots language ability for different age-groups, and using these grey block visualisation we can stack a each of the age groups.

Here we can see that generally across the age-groups the majority of people don’t understand spoken Scots, but there is a “bulge” around those aged 65 - 69 where there are more Scots writers and smaller proportion of people who don’t understand Scots.

The age range from 15 to 49 is pretty consistent with 22% being able to speak, read and write Scots. After the 65-69 “bulge” the older people - aged 75 and over - return to this 22% level.

Its kind of cute that the 3 to 4 age group can’t write Scots and only a tiny fraction can read Scots. We might even speculate that below school-age these children represent Scots-speaking households.

Council areas and Scots language ability

A similar visualisation can be done for the different council areas. This is data that has been available from the census since May 2024.

Here we can see some council regions have an above average proportion of people with Scots language skills. These regions include northern areas - Aberdeenshire, Moray and Angus, the Northern Isles - Orkney and Shetland, but also central regions - Fife, Clackmannanshire and Ayrshire.

There are a wide tranche of regions with close to average Scots language skills - Lothian and Lanarkshire.

And there are regions with distinctly below average Scots language skills - we might expect Gaelic speaking areas such as Na h-Eileanan Siar, the Highlands and Argyll & Bute, but also the populous cities - Glasgow and Edinburgh, have below average Scots skills.

Its Industrial

The census flexible table-builder that was launched at the end of November 2024, allows us to get similar grouped data for the various industrial sectors.

Its all a bit dry as just numbers in a table, so we can use whatever spreadsheet tools we have to hand to visualise it.

Here we can see that some industries have above average proportions of Scots language skills - Construction, Transport and Storage, Manufacturing, and Agriculture are the largest of these Scots speaking industries.

Industries that require more qualifications to get into seem to have a lower than average proportions of people with Scots language skills - Education, Finance, Science, and Information and Communications.

Highest level of education

There was a question in the census that asked the highest level of qualification

There are interesting distinctions between each qualification class and Scots language ability, which become clear when we look at the visualisation.

I would theorise from this that schools are actively filtering out people from Scots-speaking households. Rather than “educating the Scots out of people”, the system with its lack of Scots language teachers and teaching materials looks like it isn’t the place for Scots speakers and historically they progressively get the message and “the Scots people drop out of education”.

The usual progression of lower school - upper school - further education - university saw progressively fewer and fewer Scots speakers.

The route of HNDs and apprenticeships sees progressively more Scots speakers.

There’s all sorts of societal implications here. Deprived of university qualifications means that Scots speakers are disproportionately excluded from higher paying jobs and industries compared to people outwith Scots-speaking households.

Also there’s probably a whole untapped pool of Scot-speakers with the intelligence and aptitude to have degree-level qualification and high-paying jobs, if only the education system catered for them. Qualifying and employing them appropriately might add around £8 billion to Scottish GDP.

Luckily this is beginning to change, there are large number of teachers getting trained up to use Scots more often in the classroom. Children are flocking to take the Scots Language and Culture SQA courses when schools offer it.

Maybe in ten years’ time the census will show some changes. Its going to take the best part of a lifetime for all of society to change, but it might be starting soon.

How to

Whilst I’ve played around with the census data and the flexible table-builder for a few weeks now, it might not be intuitive for everyone, so here’s something of a step-by-step guide to grabbing data.

First clickety-click on this link to get you to the Scotland’s Census website

https://www.scotlandscensus.gov.uk/

It should look like this and you can click through to “Search census results”

From there you can click through to “Build a table”

It takes a while for the page to load completely, there’s a handful of generic tables in the middle window which might be useful, but I’ve never played with the. When its all loaded, you can choose whether you want data about families, households or individual people. I have always selected “Person” in the “Large census geographies” folder, the other options might be fun.

Then click the “New Table” button at the bottom.

Because this Substack is mostly concerned with Scots language stuff, we will select that data first. Scroll down the list of fields until you get to “Scots language skills - 8 groups, all” and tick all the boxes.

I mean you could choose some other dataset if there was something else you might want to look at.

After you’ve ticked the boxes, click on the red “Column” button, I don’t know what the wee down arrow does. It might be useful later.

Now try to find some other dataset to cross reference with the Scots language skills. Here I will select “General health” and tick all the boxes.

Then click on the “Row” button.

On the left we will now see a big empty table with the columns set to each Scots language skill category and the rows set to be each General health category.

If we click the “Retrieve Data” button, it will quickly grab the data that we want.

And then displays them as a table.

This table can either be copy and pasted, or downloaded, and opened in Excel where you can play with it to your heart’s content.

Like me, you can find weird correlations, create graphs and visualisations and share them online.

If you find any interesting Scots language-related things, please let me know.