def __init__(self,
key,
controls_csv,
tazset=None,
puma_data_dir=None,
fips_file=None,
write_households_csv=None,
write_persons_csv=None,
write_append=False,
start_hhid=1,
start_persid=1):
pd.options.display.width = 200
pd.options.display.float_format = '{:,.3f}'.format
pd.options.display.max_columns = 30
# start ids here
self.start_hhid = start_hhid
self.start_persid = start_persid
# Starter.__init__(self, key, '06', '075')
self.c = Census(key, base_url=puma_data_dir, fips_url=fips_file)
self.hh_csvfile = None
if write_households_csv:
self.hh_csvfile = open(write_households_csv,
'a' if write_append else 'w')
self.per_csvfile = None
if write_persons_csv:
self.per_csvfile = open(write_persons_csv,
'a' if write_append else 'w')
# if appending, no header
if write_append:
self.wrote_hh_header = True
self.wrote_pers_header = True
# Read the control file
print "\n\nReading the control file [%s]" % controls_csv
self.controls = pd.read_csv(controls_csv, index_col=False)
# Limit to only the specified TAZs
if tazset and len(tazset) > 0:
print "Using only TAZs in %s" % str(tazset)
self.controls = self.controls[self.controls.SFTAZ.isin(tazset)]
self.tazToPUMA2010 = pd.read_csv(
r"Q:\Model Development\Population Synthesizer\4. Geographic Work\Census 2010 PUMAs\TAZ2454_to_Census2010PUMAs.csv",
index_col=0,
converters={'PUMA2010': str})
self.state = '06'
# for caching - indexed by puma
self.h_pums = {}
self.jd_households = {}
self.p_pums = {}
self.jd_persons = {}
def get_acs_data(county, spec, settings):
state = settings['state']
census_year = settings['census_year']
if settings['tract'] == 'None':
tract = None
else:
tract = settings['tract']
c = Census(os.environ[settings['census_key']])
hh_bg_columns = get_column_names('block_group', 'household', spec)
hh_tract_columns = get_column_names('tract', 'household', spec)
if len([x for x in hh_bg_columns if x in hh_tract_columns]) > 0:
raise RuntimeError("The same acs column is being used as block group and tract. Please check expression file.")
h_acs = c.block_group_and_tract_query(
hh_bg_columns, hh_tract_columns, state, county,
merge_columns=['tract', 'county', 'state'],
block_group_size_attr=settings['hh_bg_size_attr'],
tract_size_attr=settings['hh_tract_size_attr'],
tract=tract,
year=census_year)
pers_bg_columns = get_column_names('block_group', 'person', spec)
pers_tract_columns = get_column_names('tract', 'person', spec)
if len([x for x in pers_bg_columns if x in pers_tract_columns]) > 0:
raise RuntimeError("The same acs column is being used as block group and tract. Please check expression file.")
p_acs = c.block_group_and_tract_query(
pers_bg_columns, pers_tract_columns, state, county,
merge_columns=['tract', 'county', 'state'],
block_group_size_attr=settings['pers_bg_size_attr'],
tract_size_attr=settings['pers_tract_size_attr'],
tract=tract,
year = census_year)
all_acs = h_acs.merge(p_acs, how = 'left', on = ['state', 'county', 'tract', 'block group'])
return all_acs
def c():
return Census('bfa6b4e541243011fab6307a31aed9e91015ba90')
state='25'
state_code='ma'
elif city=='Detroit':
state='26'
state_code='mi'
ALL_ZONES_PATH='./scripts/cities/'+city+'/clean/model_area.geojson'
SIM_ZONES_PATH='./scripts/cities/'+city+'/clean/sim_zones.json'
OD_PATH='./scripts/cities/'+city+'/raw/LODES/'+state_code+'_od_main_JT00_2015.csv'
ALL_SYNTH_HH_PATH='./scripts/cities/'+city+'/clean/all_synth_hh.csv'
ALL_SYNTH_PERSONS_PATH='./scripts/cities/'+city+'/clean/all_synth_persons.csv'
SIM_POP_PATH='./scripts/cities/'+city+'/clean/sim_pop.json'
VACANT_PATH='./scripts/cities/'+city+'/clean/vacant.json'
FLOATING_PATH='./scripts/cities/'+city+'/clean/floating.json'
c = Census('7a25a7624075d46f112113d33106b6648f42686a')
# load the block group geojson for the whole area
# get set of tracts covered
# identify the data we want at tract and block group level
#Households
income_columns = ['B19001_0%02dE'%i for i in range(1, 18)]
vehicle_columns = ['B08201_0%02dE'%i for i in range(1, 7)]
workers_columns = ['B08202_0%02dE'%i for i in range(1, 6)]
families_columns = ['B11001_001E', 'B11001_002E']
# year_built_columns= ['B25034_001E', 'B25034_002E', 'B25034_003E']
# includes vacant structures?
tenure_columns=['B25063_001E', 'B25075_001E']
block_group_columns = income_columns + families_columns + tenure_columns
