On 13.08.2021 22:11, Andy Jennings wrote:
That reminds me of the ternary plots for Condorcet cycles. Let
x: A>B>C
y: B>C>A
z: C>A>B
with x+y+z = 1. Then every feasible such election can be represented by
a set of barycentric coordinates and thus be plotted as a triangle.
On a related note, I imagine that you could make a more general
fingerprint for k-voter elections by placing a space-filling curve in
the appropriate high dimension space and parametrize every point as a
distance along that curve, then reduce to the same curve in say, 2D and
invert the plot to get 2D coordinates.
-km
On 13.08.2021 22:11, Andy Jennings wrote:
> I made a visualization for 3-candidate, 3-voter, ranked elections.
>
> Here is a video explaining it:
> https://youtu.be/Uvax1Hj8t_E <https://youtu.be/Uvax1Hj8t_E>
>
> Here is the website:
> https://hexagon.bettervoting.org/ <https://hexagon.bettervoting.org/>
>
> Here is the Github Repo:
> https://github.com/abjennings/socialchoice-hexagons
> <https://github.com/abjennings/socialchoice-hexagons>
>
> I had hoped to improve it all and make a version 2, but haven't had the
> time since February, so I'm letting it out there as-is.
>
> Feedback is welcome. Pull requests are also welcome. :)
That reminds me of the ternary plots for Condorcet cycles. Let
x: A>B>C
y: B>C>A
z: C>A>B
with x+y+z = 1. Then every feasible such election can be represented by
a set of barycentric coordinates and thus be plotted as a triangle.
On a related note, I imagine that you could make a more general
fingerprint for k-voter elections by placing a space-filling curve in
the appropriate high dimension space and parametrize every point as a
distance along that curve, then reduce to the same curve in say, 2D and
invert the plot to get 2D coordinates.
-km
