2014 Minnesota Budget

%matplotlib inline
import re
import pandas as pd
from requests import get

from libextract import extract
from libextract.strategies import TABULAR

# fuzzy-table-formatter branch
# https://github.com/datalib/libextract/tree/fuzzy-table-formatter
from libextract import prototypes

url = "https://www.revisor.mn.gov/laws/?year=2014&type=0&doctype=Chapter&id=294"
minn2014 = get(url)

strat = TABULAR + (prototypes.convert_table,)
tabs = list(extract(minn2014.content, strategy=strat))

table = tabs[1]
df = pd.DataFrame.from_dict(table)

#convert
def convert_num(n):
    return float(n.replace(',','').replace('(','').replace(')',''))

df['2'] = df['2'].apply(convert_num)

final_df = df[['0','2']]
final_df.columns = ['Expense Name', 'Amount']

expense_df = final_df.sort(columns='Amount')
expense_df = expense_df.set_index('Expense Name')

exclude_index = ["TOTAL", "Bond Proceeds Fund (General Fund Debt Service)"]
small_expense = expense_df[(expense_df.index != exclude_index[0]) &
                           (expense_df.index != exclude_index[1])]

pd.options.display.mpl_style = 'default'
small_expense.plot(kind='barh', figsize=[6, 10],
                   title="2014 MN Capital Budget Spending")