In [1]:

    

%matplotlib inline
import matplotlib
import matplotlib.pylab as plt
import seaborn as sns
matplotlib.rcParams['savefig.dpi'] = 256


    

In [2]:

    

from datetime import datetime
print "results as of ", datetime.now().strftime("%c")


    

    




results as of  Wed Nov  9 00:43:24 2016

In [3]:

    

from functools import partial
import requests
import pandas as pd

def num_votes(name, candidates):
    return sum([x['n'] for x in candidates if x['name'] == name])

def by_state(state):
    results = requests.get('http://elections.huffingtonpost.com/2016/results/state/{}.json'.format(state))
    data = results.json()

    geos = data['president']['geos']
    df = pd.DataFrame(geos)


    df['trump'] = df['candidates'].apply(partial(num_votes, 'Trump'))
    df['clinton'] = df['candidates'].apply(partial(num_votes, 'Clinton'))

    df2 = df[df['fractionReporting'] != 0.]

    result = pd.DataFrame({
        'count': df2[['trump', 'clinton']].sum(),
        'projection': df2[['trump', 'clinton']].divide(df['fractionReporting'], axis=0).sum()
    })
    
    return result / result.sum()

contested_states = ['AZ', 'MN', 'WI', 'MI', 'PA', 'NH']

dfs = {s: by_state(s) for s in contested_states}


    

In [4]:

    

fg, axes = plt.subplots(3, 2, sharex=True, sharey=True)
for (state, df), ax in zip(dfs.items(), axes.reshape([6])):
    title = "{}: T: {:.1f}% C: {:.1f}%".format(state, 100 * df['projection']['trump'], 100 * df['projection']['clinton'])
    df['projection'].plot(kind='bar', color='rb', ax=ax, alpha=.8, title=title)
plt.savefig('image.jpeg')