How to Use EDA to Study the Relationship Between Income and Happiness

Understanding the relationship between income and happiness is a key area of interest in both economics and psychology. Exploratory Data Analysis (EDA) provides a powerful approach for uncovering patterns, trends, and correlations in data without making prior assumptions. By using EDA, analysts can assess how income levels relate to reported happiness scores, identify anomalies, and prepare for more advanced modeling or statistical testing. Here’s a step-by-step guide on how to use EDA to study the relationship between income and happiness.

Understanding the Variables

Before performing EDA, it’s crucial to define the variables:

• Income: Typically measured as personal or household income, either in actual currency or adjusted for purchasing power parity.

• Happiness: Often quantified using survey data (e.g., a Likert scale from 1 to 10), where respondents rate their overall life satisfaction or emotional well-being.

These two variables can be influenced by many external factors such as age, education, employment status, and geographic location, all of which can be explored during the analysis.

Step 1: Data Collection and Cleaning

The first step is acquiring a dataset that includes both income and happiness data. Popular sources include:

• World Happiness Report

• Gallup World Poll

• OECD Better Life Index

• National household surveys

After obtaining the dataset:

1. Check for Missing Values: Use .isnull().sum() in pandas to assess missing data.

2. Handle Missing Data: Apply techniques like imputation or row deletion if necessary.

3. Convert Categorical Variables: For example, convert income brackets into numeric values.

4. Normalize Income Data: If income is in different currencies or units, convert it into a common scale (e.g., USD PPP).

Step 2: Descriptive Statistics

Start your analysis by summarizing the data:

• Mean, Median, Mode: Understand the central tendency of income and happiness.

• Standard Deviation and Range: Measure the spread of the data.

• Distribution Plots: Use histograms or KDE plots to visualize the distribution of both variables.

python
CopyEdit
import seaborn as sns
import matplotlib.pyplot as plt

sns.histplot(df['income'], kde=True)
plt.title('Income Distribution')
plt.show()

sns.histplot(df['happiness_score'], kde=True)
plt.title('Happiness Score Distribution')
plt.show()

Step 3: Univariate Analysis

Explore each variable individually to understand their properties.

• Income: Identify if the distribution is skewed (common in income data).

• Happiness: See if happiness scores are normally distributed or skewed.

You may also want to look at outliers using boxplots, which can reveal unusually high or low incomes or happiness ratings.

Step 4: Bivariate Analysis

To examine the relationship between income and happiness:

1. Scatter Plot: Plot income against happiness to visualize correlation.

python
CopyEdit
sns.scatterplot(x='income', y='happiness_score', data=df)
plt.title('Income vs Happiness')
plt.show()

2. Correlation Coefficient: Calculate Pearson or Spearman correlation depending on data distribution.

python
CopyEdit
df[['income', 'happiness_score']].corr(method='pearson')

A positive coefficient indicates a direct relationship, while a value near zero implies weak or no linear relationship.

Step 5: Income Brackets and Happiness

Income may not relate to happiness linearly. It helps to categorize income into brackets:

• Low income

• Middle income

• High income

Then, use boxplots or bar charts to compare happiness scores across these groups.

python
CopyEdit
df['income_bracket'] = pd.qcut(df['income'], q=3, labels=['Low', 'Medium', 'High'])
sns.boxplot(x='income_bracket', y='happiness_score', data=df)
plt.title('Happiness Score by Income Bracket')
plt.show()

This helps identify diminishing returns—i.e., whether increased income leads to significantly higher happiness only up to a point.

Step 6: Multivariate Analysis

Include other variables that could affect the income-happiness relationship:

• Age

• Education

• Marital Status

• Employment

Use pair plots or heatmaps to visualize interactions.

python
CopyEdit
sns.pairplot(df, vars=['income', 'happiness_score', 'age'])
plt.show()

sns.heatmap(df.corr(), annot=True, cmap='coolwarm')
plt.title('Correlation Matrix')
plt.show()

This provides context and helps uncover whether the correlation holds when accounting for other factors.

Step 7: Geographical and Cultural Breakdown

The income-happiness relationship can vary by region or culture. Analyze subgroups by country or region:

python
CopyEdit
sns.boxplot(x='country', y='happiness_score', data=df)
plt.xticks(rotation=90)
plt.title('Happiness by Country')
plt.show()

This can reveal if certain countries show stronger correlations or if cultural norms affect how people perceive happiness relative to income.

Step 8: Time Series and Trends

If the dataset contains data over several years, observe how the relationship evolves:

• Does rising national income correlate with rising happiness?

• Are there external shocks (like economic crises) that affect the relationship?

python
CopyEdit
sns.lineplot(x='year', y='income', data=df, label='Income')
sns.lineplot(x='year', y='happiness_score', data=df, label='Happiness')
plt.title('Income and Happiness Over Time')
plt.legend()
plt.show()

Step 9: Regression Analysis (Optional)

Although not strictly part of EDA, a simple linear regression can provide further insight:

python
CopyEdit
import statsmodels.api as sm

X = sm.add_constant(df['income'])
model = sm.OLS(df['happiness_score'], X).fit()
print(model.summary())

This gives you a model to quantify the impact of income on happiness and the statistical significance of the relationship.

Step 10: Key Insights and Interpretation

After conducting EDA, summarize the findings:

• Is there a clear relationship between income and happiness?

• Does this relationship differ across income levels or countries?

• What variables modify or mediate this relationship?

• Are there diminishing returns at higher income levels?

Insights like these can inform public policy, business decisions, and psychological studies.

Conclusion

EDA offers a comprehensive and visual approach to exploring how income affects happiness. By examining distributions, correlations, and contextual factors, analysts can uncover nuanced relationships and build a data-driven foundation for deeper statistical modeling. While income often plays a role in happiness, EDA helps to illustrate that the relationship is complex and influenced by multiple dimensions beyond financial wealth.