Ever wondered what a political map of Australia looks like at booth level? We can take the GPS positions of the polling places at the last election and plot them over a map of the federal electoral boundaries for 2007. They’re colour coded so that the darker the red, the higher the ALP two party preferred, the darker the blue the higher the Coalition two party preferred and the yellow booths are the marginal booths where the ALP TPP was between 48.8% and 51.3%. Also, the independents are in the blue column. I’ve also colour coded the electorates – light pink for Labor held and light blue for non-Labor held. Just click on the images to expand them.
South East QLD and NSW
Victoria and WA
SA and South East Australia
29 Comments
Fascinating distribution of… errr… the pink bits. It’s almost like the eastern strip and tassie are a whole different country. Any geographical distribution parallels with the united states I’m wondering? Is there some sort of “coast effect” where living closer to the sea effects your likelihood of voting in a certain way?
Well, very few people live away from the coast. So while it’s true that inland voters are more conservative, there’s a lot more conservative voters on the coast than inland.
Ben, did you have any luck exporting the mapinfo files to KML?
I’m actually thinking of exporting all my GIS data across to google earth (which is what this map is the first stage of) – any good KML software tools you recommend?
Nice.
But something is amiss in south-west Sydney. Macarthur does not have that weird a shape. And you appear to have invented a boothless electorate next door.
David, the boundaries are the ABS and AEC 2007 electoral boundary ESRI shapefiles.
Some of the booths arent shown at the level of resolution that those charts were snapped – I have to zoom in even further for them to pop us visually, which is a quirky thing that I can’t really be bothered to fix up since I’m going to eventually port all this over to google earth where it won’t be an issue.
Also some of the lat/long is out by a small amount because of projection issues – not that it should make any difference at the scales used here.
That part of Macarthur that looks like it’s been split off actually has very few people living in it. It’s possible that it’s accurate to have no polling booths in those areas.
I did manage to use ogr2gui to convert the CCD maps for NSW into KML but the file becomes freaking massive.
So I’ve started opening it bit by bit and saving them as digestible files for each LGA. But I was originally trying to use it to get a submission in for the redistribution for Friday, but that won’t work.
I’ve also in the past used this extension to mapwindow: http://interactiveearth.blogspot.com/
This hasn’t always worked. I could never work out how to convert the Canadian federal electorates into KML, so I gradually redid the maps from scratch. Clearly that’s not ideal.
I haven’t been able to find the free GIS files for the UK electorates, otherwise I would just use that. So gradually I’m drawing them manually.
Also, Possum, where did you get the data for all the polling booths? I’ve generally done it manually, seat by seat, but obviously that doesn’t allow you to do it on a greater scale.
Ben, latitude and longitude for each polling place at the last election:
http://results.aec.gov.au/13745/Website/Downloads/GeneralPollingPlacesDownload-13745.csv
Ah, brilliant. There’s probably a way to convert that all into KML in a couple of minutes.
I’ve already done it manually for Macarthur and the five inner-city seats, but that will be useful in future.
There’s dozens of those dedicated open source csv2kml apps around.
Cute.
What’s even cuter is I can see our booth in Indi. Beechworth and (I think) Stanley are the only Labor voting booths in Indi? Or is it Beechworth and Yackandandah?
Fiz, you’re right on Beechworth and Stanley being Labor booths – but also Lurg and Chiltern.
Occasionally Mt Beauty goes our way as well.
It was pretty close this time as well Zoomster, with the Coalition winning the booth 479 to 463 votes.
Possum, if you have Google Earth, you should be able to import ESRI Shapefiles (as long as they’ve got a defined prjection) and then save it as a KML file. I haven’t done it for a while as I go mostly the other way now, but I think it may lose the colouring information from your GIS software.
This might only be with the Pro version, I’m not sure about the general release one.
If anyone ever doubts that Canberra is an ALP town, just piont them to that cluster in Fraser. Doesn’t get much more definitive than that.
The other advantage using Google Earth, Possum, is you can turn these sorts of things into interactive maps, as I did with some of my posts on Queensland and Fremantle.
Although Google Maps doesn’t have as much flexibility regarding placemarks as Google Earth, eg. you can’t vary the size of booths and change the colours of set placemark icons.
Actually, that big red dot on Canberra isn’t just the seat of Fraser, it’s the seat of Canberra too. While Canberra looks a lot bigger on a map, almost its entire population is in the northern end of the electorate near Fraser.
Fiz-11: My step-brother, a former card-carrying ALP member, lives in a less labor voting area of Indi. He worked the Dederang booth in 2001. He had a very long, boring day.
TD-16: And yet they re-elected Captain Rubberstamp Gary Humphries.
Thanks for the info on the ALP booths in Indi Possum. I must admit that I hadn’t heard of Lurg before. I’m not sure what distinguishes it from the other farming areas around it though? Maybe a large family that are traditional Labor supporters?
DrMick – That makes me think that on election days I should travel around to the booths nearby and hand out sandwiches and chocolate to the poor Laborites
I know we had a couple of Labor members who moved to Lurg.
Given their community involvement, they may have made some converts!!
Fiz, get Poss to give you my email and we’ll talk sandwiches and chocolate.
No shock at all looking at Farrer’s booths, I’ll raise a glass to Broken Hill. Great stuff Poss, here’s my take on why the map looks so blue
Pretty close correlation to rainfall isn’t there [ignoring one or two bits]?
It’s a pity you have projected 2010 figures onto 2007 boundaries, though I presume you didn’t have access to the 2010 boundaries. The Macarthur boundary is obviously wrong, it doesn’t correspond to any boundaries Macarthur has ever had, nor does it correspond to state or LGA boundaries.
Psephos,
This thread is nigh on two years old!
It’s what we had to work with…. way back then. My Tardis was in the repair shop
Sorry – someone posted it on Facebook and I assumed it was new.