WORLD HAPPINESS ANALYSIS

Data Visualization

In this final analysis, I delved into the intricate dynamics between income inequality (measured by the Gini Index), World Happiness Score (WHS), and two potential mediators: generosity and social support. Although only generosity was plotted in the visual representation below, both factors were examined using Python's statistical modeling tools. The aim was to investigate whether the influence of income inequality on WHS varied depending on the levels of generosity and social support within nations. By unraveling these complexities, this analysis sheds light on the nuanced interplay between societal factors and subjective well-being, offering valuable insights for policymakers and researchers alike.

Generosity as a Mediator:

The OLS regression model revealed that both the Gini Index (p = 0.019) and generosity (p < 0.001) significantly influenced WHS.

The interaction term (gini:generosity) demonstrated a significant effect (p < 0.001), suggesting that the relationship between income inequality and WHS was moderated by generosity.

Social Support as a Mediator:

Similarly, the OLS regression model exploring the mediating role of social support unveiled significant associations between WHS, the Gini Index, and social support.

However, unlike generosity, the interaction term (gini:social) did not yield statistical significance (p = 0.719), indicating that social support may not mediate the relationship between income inequality and WHS as strongly.

These findings underscore the multifaceted nature of societal well-being and underscore the importance of considering diverse factors when evaluating the impact of income inequality on happiness levels.

Data Analysis

Python was utilized to generate the plot and OLS output. Click the button below to unveil the Python code.

Python [Generosity] import pandas as pd
import statsmodels.api as sm
import statsmodels.formula.api as smf
import numpy as np

# Load data
data = pd.read_csv('./data/WHR_2019_Full.csv')

data = data.dropna(subset=['Overall rank', 'Continent', 'Country or region', 'Score', 'GDP per capita', 'Social support',
'Healthy life expectancy', 'Freedom to make life choices', 'Generosity', 'Perceptions of corruption',
'Gini index'])

overall = data['Overall rank'].values
continent = data['Continent'].values
country = data['Country or region'].values
score = data['Score'].values
gdp = data['GDP per capita'].values
social = data['Social support'].values
lifestyle = data['Healthy life expectancy'].values
freedom = data['Freedom to make life choices'].values
generosity = data['Generosity'].values
perception = data['Perceptions of corruption'].values
gini = data['Gini index'].values

model = smf.ols(formula='score ~ gini + generosity + gini:generosity', data=data)

results = model.fit()
print(results.summary())

import matplotlib.pyplot as plt

# Create dummy variables for social support (low & high)
data['generosity_dummy'] = pd.get_dummies(data['Generosity']).iloc[:, 1]

# Generate predicted values for different levels of social support
pred_data = pd.DataFrame({'gini': data['Gini index'], 'generosity_dummy': data['generosity_dummy']})
pred_data['predicted_happiness'] = results.predict(pred_data)

# Scatter plot
scatter = plt.scatter(gini, score, c=generosity, cmap='plasma', edgecolors='black', linewidths=0.2, s=75)
plt.xlabel('Gini Index', fontweight='bold')
plt.ylabel('World Happiness Score', fontweight='bold')
plt.title('Mediating Effect of Generosity on Gini Index against Happiness', fontweight='bold')
cbar = plt.colorbar(scatter)
cbar.set_label('Generosity', fontweight='bold')
plt.gca().invert_xaxis()

plt.show()

Python was utilized to generate the plot and OLS output. Click the button below to unveil the Python code.

Python [Social Support] import pandas as pd
import statsmodels.api as sm
import statsmodels.formula.api as smf

# Load data
data = pd.read_csv('./data/WHR_2019_Full.csv')

data = data.dropna(subset=['Overall rank', 'Continent', 'Country or region', 'Score', 'GDP per capita', 'Social support',
'Healthy life expectancy', 'Freedom to make life choices', 'Generosity', 'Perceptions of corruption',
'Gini index'])

overall = data['Overall rank'].values
continent = data['Continent'].values
country = data['Country or region'].values
score = data['Score'].values
gdp = data['GDP per capita'].values
social = data['Social support'].values
lifestyle = data['Healthy life expectancy'].values
freedom = data['Freedom to make life choices'].values
generosity = data['Generosity'].values
perception = data['Perceptions of corruption'].values
gini = data['Gini index'].values

model = smf.ols(formula='score ~ gini + social + gini:social', data=data)

results = model.fit()
print(results.summary())

import matplotlib.pyplot as plt

# Create dummy variables for social support (low & high)
data['social_support_dummy'] = pd.get_dummies(data['Social support']).iloc[:, 1]

# Generate predicted values for different levels of social support
pred_data = pd.DataFrame({'gini': data['Gini index'], 'social_support_dummy': data['social_support_dummy']})
pred_data['predicted_happiness'] = results.predict(pred_data)

# Scatter plot
scatter = plt.scatter(gini, score, c=social, cmap='plasma', edgecolors='black', linewidths=0.2, s=75)
plt.xlabel('Gini Index', fontweight='bold')
plt.ylabel('World Happiness Score', fontweight='bold')
plt.title('Mediating Effect of Social Support on Gini Index against Happiness', fontweight='bold')
cbar = plt.colorbar(scatter)
cbar.set_label('Social Support', fontweight='bold')
plt.gca().invert_xaxis()

plt.show()