Talk Elections

Discuss with maps (I fail at putting up maps).

I'd help if I knew things about Canadian geography :P

What do you mean by geography.

Like demographics of Canada.

I'd actually be willing to take this on. Does anyone have Australian demographic data on hand? I can build the models and run the stats for it if you can point me to the data.

Welcome to the Forum Calgary Manifesto!

What would you like to know?

I'm Aussie, but my where's Albertan, so I know a little of the politics of both nations.

Thanks for the warm welcome. I've had a bunch of blog traffic come from this site, so I thought I'd take a look. I'm from western Canada, and I've done a fair amount of work on electoral geography.

I have poll by poll results/maps for Canada, and can match (roughly) results to my data set. It'd be a rough estimation, but it might be interesting to see. Is data available at the poll level for Australia?

Atlantic - Fairly tossupish/lean ALP
Quebec - ALP heartland
Ontario - Bellweather
Manitoba/Sask - Lean Liberal, I think KRudd would have won here.
Alberta - The National's base
BC - Probably fairly ALPish, but could go to the Liberals on a good night for them.

Very simplistic view, yes.

Oh good! What's your blog? (You won't have enough enough posts yet to put it as a link, but if you put spaces after the dots or something...).

We have a few threads that may interest you...

Here (https://uselectionatlas.org/FORUM/index.php?topic=98141.0) is the 2008 poll-by-poll maps for much of Canada.
Here (https://uselectionatlas.org/FORUM/index.php?topic=144685.0) are demographic maps of Canada - some are provincial riding, some are federal riding, some are Census division, etc.
Here (http://www.the506.com/elxnmaps/can2011/) is the website of one of the posters from here, who has been a bit silent of late (he did those 2008 poll maps from the earlier link). That link is for his 2011 poll-by-poll maps across Canada.
Here (https://uselectionatlas.org/FORUM/index.php?topic=120501.0) is an "Australian" demographic maps thread, but it's really just maps from the state of Victoria, based on state electorate boundaries, because I didn't get any further than that. It's based off the 2006 Census, but the 2011 results are out now, so I think I'll start a new thread and do a new round of maps, based off that.

Those links (other than the 506's link) are all listed in the "Special Threads" stickied thread on the International Elections board, just down from here. I've uploaded a swag of blank Canadian election maps to the gallery - feel free to use them as you see fit for your blog, if you like them. With the exception of the Northern Territory, Australian one I uploaded last week, they're all visible for perusal in this thread (https://uselectionatlas.org/FORUM/index.php?topic=147805.0). I'm part way through a New Brunswick provincial boundaries map, but it's incomplete. I'll start on a federal one once the redistribution has been completed - if you haven't seen it already, here (https://uselectionatlas.org/FORUM/index.php?topic=145422.0) is our discussion on the progression of the redistribution. Definitely take the time to check out Earl's blog - he has a link in some of his posts. He's a dipper, but quite unbiased in his electoral analysis.

Australia does have poll-level data available, but in an imprecise manner. Whereas in Canada you vote at your polling station, and each poll only records a few hundred votes, in Australia you can vote at any polling booth in your electorate, and booths near the boundary might be joint booths with the adjoining electorate. Most people will still vote at the nearest booth to them, however if you're off on a camping trip or something, or if you're working that day, you might vote at a booth you pass on the way out of town, or duck out on your lunchbreak, so it's a little less precise - and there are certainly no clearly defined boundaries for booths, so you can't map them quite the way you can in Canada. Our metropolitan booths are considerably larger than your polls, too - more likely to have a few thousand votes cast there, than a few hundred - which further decreases their effectiveness. There are some good sites that show booth locations and numbers, but it's still less precise. This website (http://www.tallyroom.com.au/melb2012) probably provides the best examples.

EDIT: To actually include a link to Earl's Blog (http://canadianelectionatlas.blogspot.com.au/).

Oh, you must be the Calgary Election Atlas guy. Welcome to the forum, I like your blog :)

Some very impressive Alberta 2012 poll maps of Edmonton there! (http://www.jbumstead.ca/?p=207)

Thanks for the comments, always good to hear. I'll be posting a breakdown of the senate results from Edmonton sometime in the next day or so, depending how work goes. I've always just worked with ArcGIS, and shapefiles (Academic license), so it makes the mapping process quite easy.

@Earl: I've been following your site for awhile now. Your work is fantastic. I've been meaning to send a hello.

Any chance there is demographic data available from the Australian national census? If we have that, we can identify the demographic relationships with certain parties, and just use 2006 Canadian census data to extrapolate the results. As the poll data is different for Australia, we might have to run it at the riding level, but it is still doable. It'd just be a question of building proper models to account for regional eccentricities. I've been meaning to learn about Australian politics, so I'm quite excited for the chance :p

Funny you should ask - I had an email about the further release of Census data today - in just under an hour, to be precise. I'll email you a link - the Australian Bureau of Statistics has a good table and mapping program allowing you to build a table of whatever topic interests you, and then map it (with a wide variety of types of locations - including state and federal electorates, or down to CCD - a Census Collection District, which is a few hundred households and the smallest statistical area recognised by the ABS - you can even map a single or several electorates by CCD, which ends up looking somewhat like your poll by poll maps).

I'll PM you the link.

You'll have to bear with me as I figure this stuff out. The Australian Bureau of Statistics is showing the Census Local Area as the smallest geographic level. Is that the same general name? or do they just not release the lower levels. I know Stats Can doesn't release our CDA or CDB data.

On an unrelated note, this thread has inspired me to do Canada with French parties, to combine my two fields of expertise (so to say). The results in Quebec will be quite interesting (to say the least).

Btw, Smid, I'd be randomly interested in having a link to something which allows me to quickly map out some Australian census data at a fairly micro (census tract?) level.

If you don't mind my asking Sharif, could you point me in the direction of the poll level data for France? That's another one that's been high on my list to track down

The best place to get it is the Australian Bureau of Statistics Table Builder website (http://www.abs.gov.au/websitedbs/censushome.nsf/home/tablebuilder?opendocument&navpos=240). The professional version costs a bit, but the basic version is free (the basic version seems to lag behind in release times and a few other subtle differences, but is still perfectly useful from our "hobby" persepective).

I've asked Al to add it to the original post in the external links (https://uselectionatlas.org/FORUM/index.php?topic=17813.msg3350714#msg3350714) thread.

I have a few other resources that I've compiled (PDF reports and Excel spreadsheets) - I could probably email you a few of the more useful ones, Hashemite.


Here (https://uselectionatlas.org/FORUM/index.php?topic=154738.msg3345785#msg3345785) is a very impressive 2007 map.

France doesn't publicly release precinct data (it would be a good start if they bothered releasing legislative election results before 1997 in the public domain, so...), though if you go on the websites of some of the bigger municipalities some of them might have results (patchy and hard to figure out). If you want micro-level data for the large municipalities, in the absence of precinct data, you can hunt down results by 'electoral canton' (not the 'cantons Insee' which don't have infra-urban cantons) which isn't as good as precinct results (by far) but is the next best thing. I put my hands on the 2012 presidential results by electoral canton, but that's it. There is a courageous and welcome attempt by SciencesPo and the Cevipof to map out the results at a precinct level (see here: http://www.cartelec.net/) for all major towns but they don't have 2012 presidentials yet.

Of course, many of the rural towns are so tiny that it doubles up as precincts ;)

Do you have any data for Argentina or Brazil? I'd be curious to start playing with more international elections. I've only really done Canada and the US.

Unfortunately, my Spanish/Portuguese  aren't good enough to navigate their websites