Congressional redistricting is dividing up a state into an assigned number of districts, which will each hold independent elections and send the plurality winner to Washington.  Since districts are composed of Census blocks, it’s exactly rephraseable as a graph partition problem on a large but finite graph.  Unfair or “gerrymandered” districts are bad in some way (though opinions differ on exactly what way).  So can’t we just agree on the fairness metric and optimize?  In my talks I’ll explain some optimization perspectives on redistricting and I’ll argue for where the interesting action is, including some open and approachable real-world problems for you to explore.  I’ll talk about joint work with many collaborators in my lab, with an especially large role played by Amariah Becker.