This is a great UI, but the quality of the results will depend on the selection bias of the inputs. For example, if you used rally speeches for one side, vs pre

This is a great UI, but the quality of the results will depend on the selection bias of the inputs. For example, if you used rally speeches for one side, vs prepared speeches for the other side, that would skew things quite a bit. Can you elaborate on the corpus which is backing this? I've been working on something similar, but it's got less to say because we've constrained ourselves to just the presidential debates, since they're "fair". https://mytake.org/search?q=supreme%20court https://mytake.org/search?q=supreme%20court I'm a big fan of "provably neutral" project to try to bring reality into politics, so I hope your project is successful!