Talking about the urban-vs-rural divide in Canada and the US with my roommate yesterday; today I realized I could actually demonstrate it.
What are you seeing here? Each point is a single riding. The x-axis shows population density. The y-axis shows the votes for that party as a percentage of the votes to the four major national parties (Liberal, Conservative, NDP, Green), and the scale is different for each. The Bloc Quebecois are not considered here, since they only participate in Quebec.
As you can see, this data seems to show that the Liberals weakly correlate with high population density, and the Conservatives correlate strongly with low population density. I tried making similar graphs using the logarithm of population density, but that didn’t reveal anything differently.
There’s quite a lot more that can be done with the data I used, you can find it all in this Excel file.