EEO-AGI Scotland Seminar Review : Cecilia MacIntyre : 2nd Oct

Cecilia MacIntyre presented her talk on Scotland’s Census 2011: How National Records of Scotland are making the most of data. Few data gathering exercises of human statistics can rival national censuses for scale, attention to detail and accuracy. Unsurprisingly, wanting to know the kind of data collected in a census is not solely a recent ambition. Governments and leaders have always been interested in knowing whom it is they govern. In fact, an attempt at gathering this kind of data goes back almost 1000 years, to William the Conqueror’s commissioning of the famous Domesday Book. The UK has had a census every 10 years since 1801, with the exception of 1941 (due to the Second World War), and last Friday we were lucky enough to be visited by Cecilia Macintyre to learn a little about Scotland’s census data from 2011 and what it can offer us today.

The 2011 census was the first web-census and consisted of 50 questions. It includes data on origin-destination (flow) statistics to track alternative population statistics like home-to-work travel and where to identify the spatial location of ‘daytime’ population. The results of the census are being released in three phases with accompanying statistical bulletins and a highlight of the data’s key points. This pro-active approach of data release is to facilitate access to data, one of Scotland’s Census and the National Records of Scotland stated objectives.

An example of three reports Scotland’s Census have recently produced includes reports on: Gaelic Language, The inhabited islands of Scotland’s west coast and Household composition by population groups. Emphasis was made that anonymity is kept by keeping the smallest set of data at the scale of at least 40 households, which is referred to as an Output Area (OA).

Scotland’s Census is keen to have the data used in valuable and innovative ways by people outside the organization and recently ran a pilot study with statisticians from Scottish Government on equality statistics. The study revealed that, for instance ‘…people of Polish ethnicity in Scotland were the most economically active…’ and yet were also ‘…likely to be in lower skilled employment…’. The results ended up widely discussed in the UK press.

Cecilia highlighted some of the places and in what form data can be found. Below are the links to the major resources highlighted: – Scotland’s Census main page – The table index – Upcoming events and workshops – Understanding Scottish Places: Visualises population data – Datashine: A contextual and highly visual way of portraying census data visually – Know Fife Dataset: Showing thematic map data in Fife

Though the 2021 census is still 6 years away, Scotland’s Census are already beginning to consider what questions will be asked. Consultation with users of the census data will begin this month! A lot of great questions can be asked of census data, so perhaps we all ought to start considering…what might be a valuable question for the census to ask the people of Scotland?

Trevor Draeseke
(MSc in GIS at the University of Edinburgh)

O Canada O Canada

My Canadian friends are celebrating the Liberal landslide in last night’s election – in fact most of them are probably still asleep.  From watching the coverage, I have two geographical reflections, posted from a position of ignorance, so please correct me if I am wrong.  Perhaps with the bias of the British electoral system, I found it bizarre that the Canadian media were broadcasting their predicted results within minutes of the polls closing in Atlantic Canada.  Nothing wrong with that you would think, except that the polls remained  open in British Columbia for another four hours – such are the joys of a vast country which spans multiple time-zones.  This means that those in the west knew how those in the east had voted and surely this must affect their voting habits – Liberals would stay at home thinking it a ‘done deal’ while Conservatives would be running to the polling stations to try to change the developing situation.  That’s illegal in Britain – and apparently it used to be in Canada.

Second, was the process of declaring results on the basis of counts from very few polling stations within a “riding” (constituency).  Hopefully using some complex algorithm, previous performance and a magic wand, candidates were declared as “elected” when only a few thousand votes has been counted from sometimes less than a quarter of the polling stations.  I presume the geographical diversity of the constituencies was taken into account, and certainly the final result seemed to reflect these early estimates, so the process must have worked.  I did cringe however when the presenter said xxxx was elected and then occasionalyl qualify it with “well they should be when all the votes are counted”. Interesting!

In the example below, Nathan Cullen was declared “elected” on 2,800 votes with only the votes from 26 of 219 polling stations actually counted.
Example Result