# Set the width to display
pd.set_option('display.width', 120)
# Increase the number of rows to display
pd.set_option('display.max_rows', 60) 

# Get the data
happiness = pd.read_csv('data/happiness_2018.csv')

# Column map to rename
cols_dict = {'Country':'country',
             'Country or region':'country',
             'Region':'region',
             'Happiness Rank':'rank',
             'Happiness.Rank':'rank',
             'Overall rank':'rank',
             'Happiness Score':'score',
             'Happiness.Score':'score',
             'Score':'score',
             'Economy (GDP per Capita)':'gdp_per_capita',
             'Economy..GDP.per.Capita.':'gdp_per_capita',
             'GDP per capita':'gdp_per_capita',
             'Family':'family',
             'Freedom':'freedom',
             'Freedom to make life choices':'freedom',
             'Generosity':'generosity',
             'Health (Life Expectancy)':'life_expectancy',
             'Health..Life.Expectancy.':'life_expectancy',
             'Healthy life expectancy':'life_expectancy',
             'Perceptions of corruption':'trust_corruption',
             'Trust (Government Corruption)':'trust_corruption',
             'Trust..Government.Corruption.':'trust_corruption',
             'Social support':'social_support',
             'Dystopia Residual':'dystopia_residual',
             'Dystopia.Residual':'dystopia_residual',
             'Standard Error':'standard_error',
             'Upper Confidence Interval':'whisker_high',
             'Whisker.high':'whisker_high',
             'Lower Confidence Interval':'whisker_low',
             'Whisker.low':'whisker_low'
            }

# Rename the columns
happiness.rename(columns=cols_dict, inplace=True)

print(happiness.columns) # check the new column names
happiness.head() # check the values

happiness.info()

# Duplicated
print('Duplicated: {}'.format(happiness.duplicated(subset='country').sum()))

cia_files = glob('data/cia.age.*.2018.txt')
cia = pd.DataFrame()

for file in cia_files:
    c = pd.read_csv(file,
                    engine='python', sep=r'\s{3,}', header=None,
                    names=['country_cia', file.split('.')[2], 'data_year'],
                    squeeze=False, skiprows=1, index_col=0,
                    thousands=',', dtype={file.split('.')[2]:'float64'}
                   )[['country_cia', file.split('.')[2]]] # read the file
    if cia.size == 0:
        cia = cia.append(c)
        print('Initialize {}: {}'.format(file.split('.')[2], cia.shape[0])) # for the first file
    else:
        cia = cia.merge(c, on='country_cia', how='outer')
        print('Merge {}: {}'.format(file.split('.')[2], cia.shape[0]))

cia.reset_index()

cia.info()
cia

cia.describe()

# Print the countries with the min life expectancy at birth
print('TOP5 countries with the min life expectancy at birth')
cia[['country_cia', 'life_expectancy_at_birth']].sort_values(by='life_expectancy_at_birth', ascending=True).head()

# Print the countries with the min median age
print('TOP5 countries with the min median age')
cia[['country_cia', 'median_age']].sort_values(by='median_age', ascending=True).head()

# Print the countries with the max life expectancy at birth
print('TOP5 countries with the max life expectancy at birth')
cia[['country_cia', 'life_expectancy_at_birth']].sort_values(by='life_expectancy_at_birth', ascending=False).head()

# Print the countries with the max median age
print('TOP5 countries with the max median age')
cia[['country_cia', 'median_age']].sort_values(by='median_age', ascending=False).head()

happiness_cia = happiness.merge(cia, on='country', how='outer')[['country', 'score', 'population_growth']]

pd.set_option('display.max_rows', 100) # increase the number of rows to display
happiness_cia[happiness_cia.isnull().any(axis=1)].sort_values(by=['score', 'country']) # the countries don't match

# Countries map to rename
country_to_rename = {'Cote d\'Ivoire':'Ivory Coast',
                     'Congo, Republic of the':'Congo (Brazzaville)',
                     'Congo, Democratic Republic of the':'Congo (Kinshasa)',
                     'Burma':'Myanmar',
                     'Korea, South':'South Korea',
                     'Czechia':'Czech Republic',
                     'Trinidad and Tobago':'Trinidad & Tobago',
                     'Korea, South':'South Korea'
                    }
# Rename the countries
cia['country'].replace(country_to_rename, inplace=True)

cia.sample(5, random_state=11) # check the values randomly

happiness_cia = happiness.merge(cia, on='country', how='left').copy()

happiness_cia.info()

happiness_cia.sort_values(by='median_age', ascending=False)

# Select the columns of interest
cols_corr = ['country', 'score',
             'life_expectancy_at_birth',
             'median_age', 'birth',
             'population_growth', 'death'
            ]
happiness_cia = happiness_cia[cols_corr]
happiness_cia

# Get correlation matrix
happiness_cia_corr = happiness_cia.corr()
happiness_cia_corr

# Color palette for the data
palette = 'summer_r'

# Inscriptions
title = """The Relationship Between Life Expectancy at Birth And Happiness"""
description = """
Correlation of the life expectancy at birth with the happiness score by country based on 2018 data.
Data: Gallup World Poll - www.kaggle.com/unsdsn/world-happiness & CIA - www.cia.gov/library/publications/the-world-factbook | Author: @data.sugar
"""

# Plot size
figsize = (6,4)

# Set the figure
sns.set(context='paper', style='ticks', palette=snark_palette,
        rc={'xtick.major.size': 4, 'ytick.major.size':4,
            'axes.spines.left': True, 'axes.spines.bottom': True,
            'axes.spines.right': False, 'axes.spines.top': False
           }
       )

# Create the plot
fig = plt.figure(figsize=figsize, facecolor='w')
ax = sns.scatterplot(x='life_expectancy_at_birth', y='score',
                     hue=happiness_cia['score'].tolist(),
                     size=happiness_cia['life_expectancy_at_birth'].tolist(),
                     sizes=(10,100),
                     data=happiness_cia,
                     palette=palette, legend=False
                    )

# Set some aesthetic params for the plot
ax.set_title(title, fontdict={'fontsize': 16}, loc='center', pad=10, c=snark_palette[-1]) # set a title of the plot
ax.annotate(description, xy=(0.1, -0.015), size=6, xycoords='figure fraction', c=snark_palette[-1])
ax.spines['bottom'].set_linestyle((0, (1, 10)))
ax.spines['bottom'].set_color(snark_palette[-1])
ax.spines['left'].set_linestyle((0, (1, 10)))
ax.spines['left'].set_color(snark_palette[-1])
ax.set_xlabel('Life Expectancy at Birth', loc='center', size='x-large', c=snark_palette[-1]) # set label of x axis
ax.set_xticks([i for i in range(50, 100, 10)])
ax.set_xticklabels([i for i in range(50, 100, 10)], c=snark_palette[-1])
ax.set_ylabel('Score', loc='center', size='x-large', c=snark_palette[-1]) # set label of y axis
ax.set_yticks([i for i in range(2, 9)])
ax.set_yticklabels([i for i in range(2, 9)], c=snark_palette[-1])
ax.tick_params(axis='both', labelsize='small', colors=snark_palette[-1], direction='out') # set x/y ticks

# Save and plot
plt.savefig('plot.pic/plot.happiness.life_exp_at_birth.png', dpi=150, bbox_inches='tight')
plt.show()

# Calculate min and max median age
median_age_min = cia['median_age'].min()
median_age_max = cia['median_age'].max()

# Calculate bins
bin_step = 5
bin_from = int((median_age_min // bin_step) * bin_step)
bin_to = int((median_age_max // bin_step) * bin_step + bin_step)
bins = [i for i in range(bin_from, bin_to, bin_step)] # age groups

happiness_cia['age_range'] = pd.cut(happiness_cia['median_age'], bins)
happiness_cia_grouped = (happiness_cia[['age_range', 'median_age', 'score']].groupby('age_range')
                                                                            .mean()
                                                                            .reset_index()
                        )
happiness_cia_grouped

# Inscriptions
title = """The Relationship Between Median Age And Happiness"""
description = """
Correlation of the median age with the happiness score by country based on 2018 data.
Data: Gallup World Poll - www.kaggle.com/unsdsn/world-happiness & CIA - www.cia.gov/library/publications/the-world-factbook | Author: @data.sugar
"""

# Plot size
figsize = (6,4)

# Set the figure
sns.set(context='paper', style='whitegrid', palette=snark_palette,
        rc={'xtick.major.size': 4, 'ytick.major.size':4,
            'axes.spines.left': False, 'axes.spines.bottom': False,
            'axes.spines.right': False, 'axes.spines.top': False
           }
       )

# Create the plot
fig = plt.figure(figsize=figsize, facecolor='w')
ax = sns.violinplot(x='age_range', y='score', data=happiness_cia,
                    inner='quart', linewidth=1,
                    palette='spring_r'
                   )

# Set some aesthetic params for the plot
ax.set_title(title, fontdict={'fontsize': 16}, loc='center', pad=10, c=snark_palette[-1]) # set a title of the plot
ax.annotate(description, xy=(0.03, 0), size=6, xycoords='figure fraction', c=snark_palette[-1])
ax.xaxis.set_label_text('') # remove label of x axis
ax.text(s='Median Age', x=3, y=2.2, horizontalalignment='center', verticalalignment='center', size='large', c=snark_palette[-1]) # set label of x axis
ax.set_ylabel('Happiness score', loc='center', size='large', c=snark_palette[-1]) # set label of y axis
ax.tick_params(axis='both', labelsize='medium', colors=snark_palette[-1]) # set x/y ticks

# Save and plot
plt.savefig('plot.pic/plot.happiness.median_age.png', dpi=150, bbox_inches='tight')
plt.show()

# Color palette for the data
palette = [snark_palette[0], # red
           'lightgrey',
           snark_palette[1] # green
          ]

# Inscriptions
title = """Relationship Between Age Indicators And The Happiness score"""
description = """
Сorrelation of age indicators with the happiness score by country based on 2018 data.
Data: Gallup World Poll - www.kaggle.com/unsdsn/world-happiness & CIA - www.cia.gov/library/publications/the-world-factbook | Author: @data.sugar
"""

# Plot size
figsize = (6,4)

# Set the figure
sns.set(context='paper', style='ticks', palette=palette,
        rc={'xtick.bottom':False, 'ytick.left':False, 
            'axes.spines.left': False, 'axes.spines.bottom': False,
            'axes.spines.right': False, 'axes.spines.top': False
           }
       )

# Create the plot
fig, ax = plt.subplots(1, 1, figsize=figsize, facecolor='w')
sns.heatmap(happiness_cia_corr, mask=mask, cmap=palette,
            vmin=-1, vmax=1, center=0,
            square=False, linewidths=.5, annot=True, fmt='.2g',
            cbar_kws={'shrink': 1, 'ticks':[], 'label':'-1 negative <- correlation -> positive +1'},
            ax=ax)

# Set some aesthetic params for the plot
ax.set_title(title, fontdict={'fontsize': 16}, loc='center', pad=10, c=snark_palette[-1]) # set a title of the plot
ax.annotate(description, xy=(20, -4), size=6, xycoords='figure points', c=snark_palette[-1])
ax.tick_params(axis='both', colors=snark_palette[-1]) # set x/y ticks
ax.set_yticklabels(ax.get_yticklabels(), rotation=0) # set rotation for y tick labels

# Save and plot
plt.savefig('plot.pic/plot.happiness.age.png', dpi=150, bbox_inches='tight')
plt.show()

# Inscriptions
title = """The Relationship Between Median Age And Birth"""
description = """
Correlation of the median age with the birth rate by country based on 2018 data.
Data: Gallup World Poll - www.kaggle.com/unsdsn/world-happiness & CIA - www.cia.gov/library/publications/the-world-factbook | Author: @data.sugar
"""

# Plot size
figsize = (6,4)

# Set the figure
sns.set(context='paper', style='ticks', palette=snark_palette,
        rc={'xtick.major.size': 4, 'ytick.major.size':4,
            'axes.spines.left': False, 'axes.spines.bottom': False,
            'axes.spines.right': False, 'axes.spines.top': False
           }
       )

# Create the plot
fig = plt.figure(figsize=figsize, facecolor='w')
g = sns.jointplot(x='median_age', y='birth', data=happiness_cia,
                  kind='reg', truncate=False, dropna=True,
                  xlim=(10, 50), ylim=(0, 50),
                  marginal_kws=dict(hist=True, bins=10),
                  color=snark_palette[0]
                 )

# Set some aesthetic params for the plot
g.ax_marg_x.set_title(title, fontdict={'fontsize': 16}, loc='center', pad=10, c=snark_palette[-1]) # set a title of the plot
g.ax_marg_x.annotate(description, xy=(0.015, -0.01), size=6, xycoords='figure fraction', c=snark_palette[-1])
g.ax_joint.set_xlabel('Median Age', loc='center', size='x-large', c=snark_palette[-1]) # set label of x axis
g.ax_joint.set_ylabel('Birth', loc='center', size='x-large', c=snark_palette[-1]) # set label of y axis
g.ax_joint.tick_params(axis='both', labelsize='large', colors=snark_palette[-1]) # set x/y ticks
g.ax_joint.spines['bottom'].set_color(snark_palette[-1]) # color x axis
g.ax_joint.spines['left'].set_color(snark_palette[-1]) # color y axis
g.ax_marg_x.tick_params(axis='x', bottom=False) # disable x margin ticks
g.ax_marg_x.spines['bottom'].set_color(snark_palette[0])
g.ax_marg_y.tick_params(axis='y', left=False) # disable y margin ticks
g.ax_marg_y.spines['left'].set_color(snark_palette[0])

# Save and plot
plt.savefig('plot.pic/plot.happiness.age.birth.png', dpi=150, bbox_inches='tight')
plt.show()

blog_post = r"""
## IN SEARCH OF 🎈HAPPINESS: 🕗WHEN?  


💪The \#happiness \#challenge on the go!  

👀Today we wondered when we are becoming happier.  
To compare population age with happiness score, we'll take the World Happiness Report from \#Kaggle, which ranks 156 countries for happiness on a 10-point scale, and use \#CIA data from The World Factbook.  

📌We will try to discover the happiest age!  

✔In the previous project \#data_sugar_happiness, we have already defined the \#TOP5 least happy countries.  
This rating includes Afghanistan and the Central African Republic.  
These countries were also included in the TOP5 countries with the lowest life expectancy at birth.  

✔Life expectancy at birth describes the overall quality of life in the country and indicates the health of the population.  
Not surprisingly, the higher the life expectancy at birth, the higher the happiness score.  

✔Let's take a look at the relationship between median age and happiness.  
It seems that society in its development reaches some point of saturation, when a person, living a longer \#life, does not become happier.  

✔Some correlations indicate that the older the \#population, the less striving (or able?) to self-reproduction:  

- the higher median age, the lower population growth;  
- the higher median age, the lower birth rate.  

📝Earlier \#data_sugar_brain, we found out that 33 is a special age.  
It is curious that the age of the population over which the happiness score does not change significantly is also about 35 years old.  
Even more interesting is that the median age of the world's population is 30 years, as we saw above.  

Well, perhaps \#humanity is in its \#prime.  
On average, of course.  

\#tobecontinued

(Interested in more details? Follow the link in bio for the entire research project!)  
.  
.  
.  
\#funtime \#probably \#datascience \#datapower \#data_sugar_happiness \#happy
\#data_know_everything_and_nothing \#linkinbio \#datajournalism \#python
"""

# Check post text length for Instagram
len(blog_post)

from IPython.display import Markdown as md
md(blog_post)