Description
An important question in some voting rights and redistricting litigation in the U.S. is whether and to what degree voting is racially polarized. We developed Python software that brings together a family of statistical methods that pair voting outcomes with demographic information to infer voting patterns. In this talk we discuss the statistical models and our implementation using scientific Python tools (most notably PyMC3 and ArviZ) and demonstrate how the software has been applied in service of informing redistricting litigation and mapmaking over the past year.