python package for calculating pff

[SPTKL]

Ideas:

1. 1 metadata file for all different variables with all information we need. e.g.

   {
   'pff_variable' : 'pop25t29',
   'acs_variable' : [
           "B01001_035",
           "B01001_011"
       ],
   'domain':'demographic',
   'base_variable':'pop_1',
   'rounding':'2'
   }

2. we should use a existing python package to pull pff data instead of using request, this would make things a lot easier

   
   from census import Census

c = Census(os.environ['API_KEY']) pd.DataFrame(c.acs5.get(('NAME', ','.join(variables)), {'for': 'block group:*','in' : 'state:36 county:081'},year=2018)))

3. we would need to create 1 master spatial lookup table or object. e.g. with both census geoid and boroct
```python
{
    'geotype':'NTA2010',
    'pff_geoname':'BK01',
    'pff_geoid':'BK01',
    'acs_geoid':''
}, 
{
    'geotype':'CT2010',
    'pff_geoname':'QN43',
    'pff_geoid':'4157102',
    'acs_geoid':'36081157102'
}

User experience

from pff import Pff
pff = Pff(api_key='XXXXXXXXX')

# if we just do a NTA level calculation, in the background it should be pulling tract level data and then aggregate to NTA level
pop25t29 = pff.calculate(pff_variable='pop25t29', geotype='NTA', year='2018')
pop25t29.head()

it should be showing the following fields: geotype, geoname, geoid, dataset, variable, c,e,m,p,z