statcheck.py

'''
    Requires a free InfoChimps account to get census data
    You'll need to enter your InfoChimps API key in the perc_black function
    '''

from BeautifulSoup import BeautifulSoup
import urllib2
from geopy import geocoders
import simplejson

G = geocoders.Google()

def geocode(address):
    '''Geocodes an address to a lat and lon'''
    try:
        place, (lat,lon) = G.geocode(address)
        return lat,lon
    except:
        return 0,0

def perc_black(lat,lon):
    '''Returns the percentage of the population that black.
       If no census data is found, returns -1.'''
    API_KEY = 'YOUR INFOCHIMPS API KEY'
    URL = 'http://api.infochimps.com/web/an/census/combined.json?apikey=' + API_KEY 
    URL = URL + "&lat=%s&long=%s" % (lat,lon)
    try:
        result = simplejson.load(urllib2.urlopen(URL))
        return float(result['percent_black'])
    except:
        return -1.0
 
states = ['AL','AK','AZ','AR','CA','CO','CT','DE','FL','GA',
          'HI','ID','IL','IN','IA','KS','KY','LA','ME','MD',
          'MA','MI','MN','MS','MO','MT','NE','NV','NH','NJ',
          'NM','NY','NC','ND','OH','OK','OR','PA','RI','SC',
          'SD','TN','TX','UT','VT','VA','WA','WV','WI','WY',]

total = 0
geocoded = 0
census = 0
black = 0

for state in states:
    url = 'http://www.plannedparenthood.org/health-center/findCenter.asp?s=%s&p=0&o=1&c=0' % (state)
    html = urllib2.urlopen(url)
    soup = BeautifulSoup(html) 
    divs = soup.findAll('div', {'class':'center vcard'})
    for div in divs:
        total += 1
        address = div.findAll('span', {'class':'center_google_address'})[0].text
        lat,lon = geocode(address)
        if lat > 0:
            geocoded += 1
            p = perc_black(lat,lon)
            if p > 0:
                census += 1
            if p > 0.5:
                black += 1

print "Total Addresses: %s" % (total)
print "Geocoded Addresses: %d" % (geocoded)
print "Addresses with census info: %s" % (census)
print "Addresses that are >50%% black: %s" % (black)