We added three Poppy File – type maps to the series for Remembrance Day:
– A (mostly) household-scale national map of about 90% of Canada’s Korean War dead, with as far as I’m aware unpublished photos of the war we bought from the National Archives;
– First World War maps based on the Canadian Great War Association data for Winnipeg and Vancouver, with about 1,000 people in each;
– Also, I had the chance to renovate the Toronto casualty maps, fixing broken links and adding more, putting the First World War, Second World War and Korea maps on a dropdown menu and adding interactive calendars like the one below:
The Vancouver map made an appearance on the Global BC morning show:
I like making ‘what-if’ maps with the Google platform. Our last one was a conceptual map showing what Canada would, or might, look like with provinces of roughly equal population. (The basic idea owed something to this map.)
The maps we published today were aimed at showing how the representation formulas in the House of Commons distort a straight rep-by-pop distribution of seats, in this case by simulating on a riding-by-riding basis what the electoral maps of Toronto and Vancouver would look like if they had ridings allocated on the same population scales as Atlantic Canada’s.
Making the math and geography come out correctly was fussy but not all that complicated. It wouldn’t have worked at all if consecutively numbered census tracts hadn’t tended to be clustered physically. What I ended up doing was breaking up census tract KML in Excel, grouping the tracts by population, seeing how the maps looked (they came out as long strips at first) then reallocating tracts between notional ridings to make them squarer, while making sure the reshaping didn’t throw the population counts off. IIRC Vancouver’s map needed seven or eight iterations of this.
When that was done, I traced the ridings in Google Earth and made a fresh KML file.
Here’s Toronto’s map:
I was working on a longer post on Toronto Life’s neighbourhood rankings project, but the Move Smartly blog beat me to it:
No data analysis is a bulletproof, perfect response to any one question. Each is best considered a way to get to the right answer. That’s why it’s important to disclose the methodology used so that readers can decide if an author or organization’s assumptions make sense to their own decision-making.
There are many strange things in the Toronto Life data, and Move Smartly mentions several of them.
Toronto Life’s ranking data measures diversity, among other things. I would very much like to see the data under which Blake-Jones scores lowest in the city, at #140, Regent Park scores #124 and Moss Park is #118. I’m not saying it’s wrong, just that it’s puzzling and that I’d like to know what is being measured and how. (In the 2011 census, tract 5350072.01, a box of: Danforth/Pape/Jones/railway, about half the residents were native English speakers; in 2006, Blake-Jones had a higher percentage of immigrants – though not recent immigrants – and a much higher number of people with no knowledge of either official language than the city as a whole.)
The whole exercise is more opaque than it should be, and it would be a much stronger package if it was more transparent.
(Also: I don’t understand how a ranking of 140 neighbourhoods by transit quality could result in: 11. Greenwood-Coxwell; 13. Danforth Village-East York; 18. Blake-Jones; 53. North Riverdale
when all four neighbourhoods are on the Danforth subway line, with North Riverdale closest to downtown, and also served by the King car on Broadview. There may be a logic, but I can’t make out what it could be.)