This page you are reading is a visualization of Tabular data about legislators. Data is based on voting records downloaded from http://leg.wa.gov See also Visualization of Washington State Senate members based on voting records for 2019.
Legislators shown close together voted similarly on motions. . Positioning is inexact and approximate, due to the stochastic algorithm used. See the bottom for details.
Click and drag with your mouse to rotate in 3D space. Use your mouse scroll bar to zoom. Click on icons right below to reposition.
Notice that Democrats (in blue) are split into two tight clusters. I see this repeatedly when I re-run the algorithm. Can someone explain why there are two groups of Dems? Centrists versus progressives? Those who vote with the caucus and those who don't?
The three scales/axes (x, y, and z) are arbitrary in this visualization and only serve to define a space on which to show the clustering of lawmakers. There's a well-developed sub-branch of AI that involves so-called "unsupervised" machine learning. Clustering is one example of unsupervised learning. The idea is to project high-dimensional data -- voting records in this case -- into lower dimensions in a way that lets you visualize and understand the data better. TSne is a recent, powerful method of clustering, and I used it to cluster the lawmakers. If I run the algorithm multiple times, details vary -- some lawmakers are moved around quite a bit -- but overall the clustering is similar.