def counties(self):
from geoid.acs import AcsGeoid
from geoid.civick import GVid, State
d = {AcsGeoid.parse(row.geoid).county_name.medium_name:str(AcsGeoid.parse(row.geoid).convert(GVid) )
for row in self.dep('geofile50') if row.state == 6}
d['California'] = State(6)
return d
def sub_geoids(v):
"""Replace state abbreviations and state fips codes with state and national geoids"""
from geoid.censusnames import stusab
from geoid.acs import Us, State, AcsGeoid
if len(v) == 2:
v = v.upper()
stmap = {v: k for k, v in stusab.items()}
if v == 'US':
return str(Us())
if v in stmap:
return str(State(stmap[v]))
# Maybe it is a state number
try:
v = int(v)
return str(State(v))
except ValueError:
pass
return str(AcsGeoid.parse(v))
def shape_url(self):
"""Return the shapefile URL"""
from geoid.acs import AcsGeoid
us = tiger_url(self.year, self.summary_level,
AcsGeoid.parse(self.geoid).stusab)
return parse_app_url(us)
def geo_url(self):
"""Return a url for the geofile for this Census file"""
from geoid.acs import AcsGeoid
us = tiger_url(self.year, self.summary_level,
AcsGeoid.parse(self.geoid).stusab)
return parse_app_url(us)
def __init__(self, ref, cache=None, working_dir=None, **kwargs):
from geoid.acs import AcsGeoid
from publicdata.census.files.metafiles import TableMeta
super().__init__(ref, cache, working_dir, **kwargs)
gid = AcsGeoid.parse(self.ref.geoid)
self.table = Table(self.ref.year, self.ref.release, gid.stusab,
str(self.ref.summary_level), self.ref.tableid)
self._meta = TableMeta(self.ref.year, self.ref.release)
assert isinstance(ref, CensusUrl)
def parse_to_gvid(v):
"""Parse an ACS Geoid or a GVID to a GVID"""
from geoid.civick import GVid
from geoid.acs import AcsGeoid
m1 = ''
try:
return GVid.parse(v)
except ValueError as e:
m1 = str(e)
try:
return AcsGeoid.parse(v).convert(GVid)
except ValueError as e:
raise ValueError("Failed to parse to either ACS or GVid: {}; {}".format(m1, str(e)))
def __init__(self, ref, cache=None, working_dir=None, **kwargs):
from geoid.acs import AcsGeoid
from publicdata.census.files.metafiles import TableMeta
super().__init__(ref, cache, working_dir, **kwargs)
gid = AcsGeoid.parse(self.ref.geoid)
try:
stusab = gid.stusab
except AttributeError:
gid.us# "US" level geoids don't have a state parameter.
stusab = "US"
self.table = Table(self.ref.year, self.ref.release, stusab,
str(self.ref.summary_level), self.ref.tableid)
self._meta = TableMeta(self.ref.year, self.ref.release)
self._source_url = kwargs.get('source_url')
assert(self._source_url)
assert isinstance(ref, CensusUrl)
def __iter__(self):
from ambry import get_library
import censuslib.dataframe
import pandas as pd
# The district NCES codes aren't in the district file, although they are in the school file.
schools = self.bundle.partition(table='schools').analysis.dataframe()
schools['cd_code'] = schools.cdscode.apply(lambda cdscode: cdscode[:7])
nces_districts = schools[schools.statustype=='Active'][['ncesdist', 'cd_code']].drop_duplicates()
# The actual districts file
ca_districts = self.bundle.partition(table='districts').analysis.dataframe()[
['cd_code', 'county_sos','county_fips','county_gvid','county','district']]
#assert len(nces_districts) == len(ca_districts)
# Combine the codes from the school file with the district file
cd_code_districts = ca_districts.set_index('cd_code').join(nces_districts.set_index('cd_code')).reset_index()
##
## Join the California state districts list with the Census districts list
###
from geoid.acs import AcsGeoid
from geoid.civick import GVid
dist_pred = lambda row: row.state ==6
def mk_cd_code(nces):
return '06{:05d}'.format(nces)
# Combine the three partitions for school districts in the census, and extract the NCES code
elem = self.bundle.dep('elementary').analysis.dataframe(dist_pred)[['geoid', 'name']].copy()
elem['nces'] = elem.geoid.apply(lambda geoid: mk_cd_code(AcsGeoid.parse(geoid).sdelm) )
second = self.bundle.dep('secondary').analysis.dataframe(dist_pred)[['geoid', 'name']].copy()
second['nces'] = second.geoid.apply(lambda geoid: mk_cd_code(AcsGeoid.parse(geoid).sdsec) )
unified = self.bundle.dep('unified').analysis.dataframe(dist_pred)[['geoid', 'name']].copy()
unified['nces'] = unified.geoid.apply(lambda geoid: mk_cd_code(AcsGeoid.parse(geoid).sduni) )
# Add a GVID
census_districts = pd.concat([elem, second, unified], axis=0)
census_districts['gvid'] = census_districts.geoid.apply(lambda geoid: AcsGeoid.parse(geoid).convert(GVid) )
# Do the join
districts = cd_code_districts.set_index('ncesdist').join(census_districts.drop_duplicates().set_index('nces')).reset_index()
districts.columns = ['ncesdist'] + list(districts.columns)[1:]
# These are different sizes, don't know why. The de-duplicated census file is smaller than the
# list from California, probably because many of the districts are smaller than the reporting limits.
# NOTE: Becase join is a left join, (a) it must be joined in the order above ( cd_code_districts, the larger list,
# is the base ) and (b) the joined 'districts' dataframe will have some missing geoids.
# The missing geoids appears to be primarily for County offices of education and districts in small counties.
# >>> print len(census_districts), len(census_districts.drop_duplicates()), len(cd_code_districts), len(districts)
# >>> 1976 988 1098 1098
districts.gvid.fillna(value='', inplace = True)
districts.geoid.fillna(value='', inplace = True)
districts.name.fillna(value='', inplace = True)
districts.ncesdist.fillna(value=0, inplace = True)
df = districts.reset_index()
yield ['id'] + list(df.columns)
for index, row in df.iterrows():
yield [index] + list(row)
