Refactor + Geography Conversion (base PR for refactor)

[SPTKL]

The 2010_to_2020 geography aggregator is start but not completed. see factfinder/geography/2010_to_2020.py for methods to implement

100

remaining issues: calculate = Calculate(api_key, 2019, 'acs', '2010_to_2020')

• since ct2020 is now considered an aggregated geography, the logic is now calculating the median age (via calculate_e_m_median) for ct2020. but it shouldn't be
• calculate('lgvielep1', 'NTA') is failing
• calculate_e_m should be recursive, so that we can do multistage calculation e.g. ct2010 -> ct2020 -> nta. however, the current set up does not meet the needs

[mgraber]

@SPTKL tract_to_nta and tract_to_cdta are now recursive within 2010_to_2020.py:

https://github.com/NYCPlanning/db-factfinder/blob/82b2b4464e672c07f2a371707f631871e9dbe06d/factfinder/geography/2010_to_2020.py#L99

https://github.com/NYCPlanning/db-factfinder/blob/82b2b4464e672c07f2a371707f631871e9dbe06d/factfinder/geography/2010_to_2020.py#L116

[SPTKL]

merging for now, further improvement will be made in separate PR