import pandas as pd
import os
curdir = os.path.abspath('./..')

df = pd.read_csv(os.path.join(curdir, 'scraped_data', 'region_all.csv'))
df[['name', 'year', 'net_profit', 'tipp', 'staff_costs', 'financial_costs', 'debt_repayments', 'allocation']].head(n=20)

df['staff_costs_per_person'] = df['staff_costs']/df['population']
df['allocation_per_person'] = df['allocation']/df['population']
df['operating_costs_per_person'] = df['operating_costs']/df['population']
df[['name', 'year', 'staff_costs_per_person', 'allocation_per_person', 'operating_costs_per_person']].head(n=26)

dfgp = df.groupby(['insee_code', 'year']).first()
dfmean = df.groupby('year').mean()
population_by_year = df.groupby('year')['population'].sum()

def plot_value(value, by_person=True):
    fig, axes = plt.subplots(figsize=(12,6))
    axes.set_title('%s by person from 2008 to 2012'%value)
    
    for reg in dfgp.index.get_level_values(0).unique()[:5]:
        years = dfgp.index.get_level_values(1).unique()
        if by_person:
            data = dfgp[value][reg]/dfgp['population'][reg]
        else:
            data = dfgp[value][reg]
        axes.plot(years, data.tolist(), label=dfgp['name'][reg][years[0]])
    if by_person:
        mean = dfmean[value]/population_by_year
    else:
        mean = dfmean[value]
    axes.plot(years, mean.tolist(), label='mean')
    axes.legend(loc='best');

plot_value('tipp')
plot_value('debt_annual_costs')p
plot_value('staff_costs')

plot_value('staff_costs', by_person=False)



df = pd.read_csv(os.path.join(curdir, 'nosdonnees', 'region_all.csv'))

df.columns.tolist()