Changes in real wages since 2007

import pandas as pd
import matplotlib.pyplot as plt

# Study the percentage real wage change between YEAR_START and YEAR_END.
YEAR_START, YEAR_END = 2007, 2015

# Read in the data to a Pandas dataframe.
df = pd.read_csv('real-wage-changes.csv', names=['LOCATION', 'INDICATOR',
        'SUBJECT', 'MEASURE', 'FREQUENCY', 'TIME', 'Value', 'Flag Codes'])

# These are the countries we're interested in, keyed by the 'LOCATION' code
# in the input file.
country_codes = {'CAN': 'Canada',
                 'CHE': 'Switzerland',
                 'DEU': 'Germany',
                 'ESP': 'Spain',
                 'FRA': 'France',
                 'GBR': 'UK',
                 'GRC': 'Greece',
                 'IRL': 'Ireland',
                 'ITA': 'Italy',
                 'JPN': 'Japan',
                 'NLD': 'Netherlands',
                 'POL': 'Poland',
                 'PRT': 'Portugal',
                 'SWE': 'Sweden',
                 'USA': 'USA'
                }

# Get the percent wage change for each country in a dictionary,
# percent_wage_change keyed by country code.
percent_wage_change = {}
for country_code in country_codes.keys():
    wage_start = df[(df['LOCATION']==country_code) & 
                    (df['TIME']==YEAR_START)]['Value'].values[0]
    wage_end = df[(df['LOCATION']==country_code) &
                    (df['TIME']==YEAR_END)]['Value'].values[0]
    wage_change = wage_end - wage_start
    percent_wage_change[country_code] = wage_change / wage_start * 100

# Get a list of country codes, sorted in increasing order of % wage change.
sorted_country_codes = sorted(percent_wage_change,key=percent_wage_change.get)

# Now plot a horizontal bar chart
fig, ax = plt.subplots(facecolor='w')
bars = ax.barh(range(len(country_codes)),
               [percent_wage_change[country_code] for country_code
                    in sorted_country_codes],
               fc='g', ec='none', align='center')

# Add labels (country names to the left of the bars, percentage real wage
# changes to the right). Also colour the bars for wage falls in red.
country_name_offset = -0.4
percent_label_offset = +0.4
for i, country_code in enumerate(sorted_country_codes):
    if percent_wage_change[country_code] < 0:
        bars[i].set_color('r')
    ax.annotate(s=country_codes[country_code],
                xy=(min(0, percent_wage_change[country_code]) +
                            country_name_offset, i),
                va='center',
                ha='right')
    ax.annotate(s='{0:.1f}%'.format(percent_wage_change[country_code]),
                xy=(max(0, percent_wage_change[country_code]) +
                            percent_label_offset, i),
                va='center',
                ha='left')

# Tidy up by removing axis spines, tick marks and tick labels.
for spine_pos in ('top', 'right', 'bottom', 'left'):
    ax.spines[spine_pos].set_visible(False)
ax.xaxis.set_ticks_position('none') 
ax.yaxis.set_ticks_position('none') 
ax.set_axisbelow(True)
ax.set_xticklabels([])
ax.set_yticklabels([])
fig.suptitle('Percentage real wage changes {}-{}'.format(YEAR_START, YEAR_END),
                fontsize=16)
ax.set_title('OECD average annual wages (PPP), 2016', fontsize=12)

plt.show()