It's been said by pretty much everyone who's commented on it that the preference flows at the 2015 Queensland election were substantially different from those in 2012 – a strong campaign by the non-LNP side to number every square helped to reduce the rate of exhaustion, increase the rate of preference flow to Labor, and decrease the rate of preference flow to the LNP.

I thought it'd be interesting to compare the exhaustion rate in each electorate in the two elections. The graph at the end of this page shows (by default) the exhaustion rate in 2015 plotted against the exhaustion rate in 2012. The regression line is

exh_2015 = 0.93*exh_2012 - 11.4,

so the exhaustion rate, previously around 50%, decreased by around 15 percentage points in this election.

But why stop there!? In the widget below (thankyou d3.js), you can set the axes to plot any of a number of variables from any Queensland election since 1992. The plot area is zoomable and pannable, and if you move your mouse over a circle, the seat name and the values for the current plot will be shown at top-left. Clicking on a circle turns it white, and you can follow that electorate as you change the axes. (For touch-screen users, mouseover and click are almost identical; the temporarily cyan circles make sense if you're using a mouse.)

We can see the correlation between the two-party-preferred vote in 1992 and that in 2015 (0.84), watch the exhaustion rate rise enormously in 2001, etc.

Some notes:

• The 2PP figures are only plotted for ALP v LNP contests.
• The percentage of preferences going to either ALP or LNP are also only plotted for ALP v LNP contests.
• For elections prior to the merger, the "LNP primary" is the sum of Liberal and National primary votes. (Or, in at least one case – Ipswich 1992 – the sum of Liberal and two National candidate votes.) Liberal and National primary vote percentages are also available separately.
• If you only change a year, then the axes remain unchanged. If you change a variable, then the zoom and axes are reset.
• The 2015 results aren't finalised yet, and the ECQ has taken away most of the two-candidate-preferred figures. I've grabbed the latest numbers from the ABC on the assumption that Antony knows what he's doing. Update: The ECQ has released the final figures for primary and two-candidate-preferred votes, and I have updated the data on this page accordingly.
• Most of the old election results also come from Antony.
• There's an annoying bug when I keep changing the axes on my phone – the JavaScript seems to forget some of the objects and throws an error. The page will reload in a few seconds if this happens.
• I hope I've stamped out all the bugs in the data wrangling itself, but there's a lot of it and plenty I haven't checked carefully.
• The circles are coloured according to the winner of the seat at a particular election. KAP and One Nation have the same colour not because I think they're the same but because I ran out of colours I was happy with, and those two parties don't clash.
• Colours are: red ALP, blue LNP/LIB, green NAT, brown KAP or ONP, dark grey for independent. A circle is transparent if the seat didn't exist in the year being used to define the colours.