Pulling back from trying to layer the weighted features on top of the churn probability. But right now feeling OK about this set-up. Where the churn probability follows a color scale from red to blue (hcl space) and the weights follow a diverging scale from white to blue and white to red.