Detailed Federal Electoral Map

[August1991]

La Presse newspaper has a detailed map using Google Earth and the 2008 federal election results: http://www.cyberpresse.ca/actualites/elections-federales/resultats-des-elections-federales-2008/

The application seems to work better with Google Chrome and you may have to download a Google Earth Plugin.

It allows you to zoom into polling station results for any riding in Canada. (It's advertised as a way to find out how your neighbours voted.)

Anyone involved in a riding campaign uses a map like this all the time. I have never seen one that is so easily accessible. To use it properly, you have to know the urban geography of the area.

[Harry Hayfield]

La Presse newspaper has a detailed map using Google Earth and the 2008 federal election results: http://www.cyberpresse.ca/actualites/elections-federales/resultats-des-elections-federales-2008/

The application seems to work better with Google Chrome and you may have to download a Google Earth Plugin.

It allows you to zoom into polling station results for any riding in Canada. (It's advertised as a way to find out how your neighbours voted.)

Anyone involved in a riding campaign uses a map like this all the time. I have never seen one that is so easily accessible. To use it properly, you have to know the urban geography of the area.

And if you Google "The Tally Room + Australian Greens" you can find a downloadbale Google Earth map showing the election results at the riding level

[Bryan]

It's really interesting to see how the economy of voting applies even within a riding. Parties have certain pockets, and even the also-ran candidates have neighborhoods where they either led or cam damn close for (that particular poll at least). In my riding at least, the CPC candidate blew out the while riding with the exception of some very small pockets that were LPC, and the NDP candidate was strong second (LPC third) in many of the CPC polls. (Do you call the actual spot on the map the "poll"?)

I wonder how hard it would be to calculate a hypothetical FPTP model in adding up the polls won (rather than votes), and see if the seat count would change. Is the Elections Canada per-polling-station data available as a spreadsheet?

[August1991]

(Do you call the actual spot on the map the "poll"?)

That's how I've always known it: A "poll" or a "poll number". According to Elections Canada, there are an average of 350 voters in a poll. With an average turnout of 60%, that means you can narrow analysis down to about 200 votes.

I wonder how hard it would be to calculate a hypothetical FPTP model in adding up the polls won (rather than votes), and see if the seat count would change. Is the Elections Canada per-polling-station data available as a spreadsheet?

That's an interesting question and presumably the data is available somewhere because this guy created the map linked by La Presse above.

I have several thoughts on your point, Bryan. First, the maps could have degrees/intensity of colour to indicate how strong the poll went to one party or the other. Second, whether FPTP or PR, one vote is just one vote and it changes nothing in terms of deciding who sits in the House of Commons. (This is a hard fact of all one-man, one-vote voting schemes that few people seem to grasp.) Third, pay attention to the turn-out rate in individual polls. I was surprised to see in certain urban ridings in poll numbers with a large immigrant population, the turn-out rate fell to as low as 30%.

----

Looking at various riding results, it's hard not to realize that people vote more often according to their self-image (or even snobism) rather than to any connection to party policy. If you know a riding well, and know who lives where, it's hard not to predict which party they will support. Party allegiance is akin to an aspirational product in marketing theory. For example, consider these two close ridings:

Papineau, Justin Trudeau's riding. Driving down its streets, you could almost predict which polls are Liberal and which are Bloc.

Parry Sound, Tony Clement's riding. The Liberal/Conservative divide is predictable.

BTW, Karl Rove became famous largely because he used data such as this to help Republicans get elected.

Edited April 5, 2011 by August1991

[Evening Star]

This is fascinating. Thanks!

Looking at various riding results, it's hard not to realize that people vote more often according to their self-image (or even snobism) rather than to any connection to party policy. If you know a riding well, and know who lives where, it's hard not to predict which party they will support. Party allegiance is akin to an aspirational product in marketing theory. For example, consider these two close ridings:

I dunno, in all the urban/suburban ON ridings I looked at, voting patterns seem to break down pretty predictably according to class/income (which I do think reflects party policy more than self-image). The NDP/CPC divisions (with no Liberal polls) in Windsor are interesting though.