Automate the comparison of service territories

[zaneselvans]

There are several contexts in which automatically identifying differences between two service territories will be useful:

• Finding cases in which using different methods to generate the territory resulted in substantial differences. (e.g. state limited vs. complete)
• Tracking how a service territory has changed from year to year.
• Comparing the "utility" geometry for an entity vs. the "balancing authority" geometry for the same entity. We want to be able apply these checks across the whole dataset to highlight discrepancies for further investigation.

New fields required in the demand_summary() table for each FERC 714 respondent:

• [x] Area of the planning area
• [x] Population of the planning area
• [x] Population density of the planning area
• [x] Per-capita demand within the planning area
• [x] Per-unit-area demand within the planning area

Year-to-year metrics of change to calculate:

• [ ] Change in total number of counties making up a territory.
• [ ] Change in the total area of the territory.
• [ ] Change in total demand reported within the territory.
• [ ] Change in population of the territory.
• [ ] Change in annual electricity demand per unit area.
• [ ] Change in annual electricity demand per person.
• [ ] Change in population density.
• [ ] Spatial difference between the two territories (computationally intensive)

Using these metrics we should be able to screen for big differences automatically, instead of having to page through hundreds of different maps, and identify cases where there's an unexplained difference.

[zaneselvans]

@ezwelty I did create these metrics, but didn't automate the inter-year comparisons to help identify bad geometries / big changes.