Jesse Richardson submitted a strong essay that articulated his goals and understanding of the issue clearly, with helpful supportive research. He did an excellent job creating a map that reflected his intentions for making competitive districts. Overall, he scored the highest in his regional cohort on most of the values metrics.
After reading about this subject, I came to the conclusion that the No. 1 goal in redistricting should be ensuring competitive elections, because of how partisan things are these days. The more competitive elections there are, the more choice voters will have and the more motivation representatives will have to be moderate and work across the aisle.
With this in mind, I drew my map such that 12 of the 18 districts have the potential to be competitive. I also prioritized equal population (this is a basic one; each vote should be worth the same) as well as party advantage. By most metrics, this map would produce a roughly even partisan split across Pennsylvania, assuming one party isn't having a strong year nationally. I didn't worry about splitting counties as these are largely arbitrary boundaries, and I also considered compactness to be largely aesthetic, and not an important consideration in our democracy.
I did find it a struggle to make competitive districts when the tool only showed registration data, but not how each local area voted. But I drew on my own knowledge and other useful sites to remedy this issue. I also found there was an inevitable and difficult trade-off between representation of certain groups and competitiveness. For example, black voters are overwhelmingly Democratic. However I think I managed to find a healthy balance, with only one district (the 18th) being unnaturally extended across multiple regions of the state.
One important point of debate is how to deal with the Democrats' intrinsic geographic disadvantage with their voters being consolidated in the city of Philadelphia. Some opinions that I have read suggested that this is no fault of the Republicans and hence shouldn't be remedied. But I think that democracy shouldn't be influenced to such a degree by geographic distribution; people vote, not land. So I did my best in drawing this map to nullify the Democratic geographic disadvantage and make a map that's fair for both parties as well as having many competitive elections. It is ultimately possible to make a map with 18 swing districts but I felt like that was a bridge too far in terms of making a messy map with districts that have no clear geographic center.