Analyzing the Determinants of Happiness (2015-2019)

Project Overview

This project analyzes the key factors influencing happiness across countries from 2015 to 2019 using data from the World Happiness Report. By examining economic, social, and health-related indicators, we aim to understand:

• The main determinants of happiness across nations.
• How these determinants change over time.
• The countries with the highest and lowest happiness scores, and what influences them.
• The correlation between economic factors (such as GDP per capita) and happiness compared to health, freedom, and social support.

This analysis provides insights that can guide policymakers in improving societal well-being.

Research Questions

1. What are the main factors influencing happiness across countries?
2. How have these factors changed over the years (2015-2019)?
3. Which countries have the highest and lowest happiness scores, and why?
4. How strong is the correlation between economic factors (GDP per capita) and happiness compared to other variables like health, freedom, and social support?

Dataset

We use the World Happiness Report datasets from Kaggle:
🔗 World Happiness Report on Kaggle

Key Variables:

• Country: Name of the country
• Region: Geographical region (for some years)
• Happiness Rank: The ranking of happiness by country
• Happiness Score: The overall happiness score
• Economy (GDP per Capita): Economic prosperity measure
• Family: Social support measure
• Health (Life Expectancy): Life expectancy metric
• Freedom: Freedom to make life choices
• Trust (Government Corruption): Perceived corruption in government
• Generosity: Donations and volunteering behavior

Project Structure

/happiness-analysis
│── /data/                     # Raw and processed datasets
│    ├── 2015.csv
│    ├── 2016.csv
│    ├── 2017.csv
│    ├── 2018.csv
│    ├── 2019.csv
│── /src/                      # Python scripts for analysis
│    ├── data_cleaning.py       # Handles missing values & fixes inconsistencies
│    ├── data_merging.py        # Merges datasets into one DataFrame
│    ├── exploratory_analysis.py # Exploratory data analysis (EDA)
│    ├── correlation_analysis.py # Correlation analysis between factors
│    ├── regression_model.py    # Regression model predicting happiness scores
│── /notebooks/                 # Jupyter Notebooks
│    ├── happiness_analysis.ipynb  # Full Jupyter notebook report
│── /docs/                      #  visualizations
│    ├── correlation_matrix.png      # Heatmap visualization
│    ├── actual_vs_predicted.png     # Scatter plot of actual vs predicted scores
│    ├── happiness_distribution.png  # Boxplot of happiness scores
│── .gitignore                  # Ignore unnecessary files
│── README.md                   # Documentation
│── LICENSE                     # Open-source license (optional)
│── requirements.txt             # Dependencies
│── Makefile                     #  Automate script execution

---

Key Visualizations

Correlation Between Happiness Factors

This heatmap visualizes the relationships between key happiness indicators.

---

Actual vs Predicted Happiness Scores

A scatter plot comparing the actual and predicted happiness scores from the regression model.

---

Happiness Score Distribution Over Time

This boxplot shows how happiness scores have changed across the years.

Installation & Setup

1️⃣ Clone the repository

git clone https://github.com/Melasery/happiness-analysis.git
cd happiness-analysis

2️⃣ Install dependencies

pip install -r requirements.txt

3️⃣ Run the analysis

Execute the scripts in order:

# 1. Data Cleaning
python src/data_cleaning.py

# 2. Merging Datasets
python src/data_merging.py

# 3. Exploratory Analysis
python src/exploratory_analysis.py

# 4. Correlation Analysis
python src/correlation_analysis.py

# 5. Regression Model
python src/regression_model.py

---

Methodology

1️⃣ Data Cleaning

• Checked for missing values and handled them appropriately.
• Standardized column names across datasets.
• Saved a cleaned dataset for consistency.

2️⃣ Merging Datasets

• Combined data from 2015 to 2019 into a single DataFrame.
• Ensured consistency across different years.

3️⃣ Exploratory Data Analysis (EDA)

• Descriptive statistics (.describe())
• Happiness score distribution visualizations
• Identification of trends over time.

4️⃣ Correlation Analysis

• Calculated the correlation matrix to measure relationships.
• Generated a heatmap to visualize strong/weak correlations.

5️⃣ Predictive Modeling

• Built a linear regression model to predict happiness scores.
• Evaluated performance using:
  • Mean Squared Error (MSE)
  • R-squared (R²) Score
• Interpreted feature importance from regression coefficients.

Results & Key Insights

Strong Predictors of Happiness

• GDP per capita, social support, and health were the top predictors of happiness.
• Trust in government had a weaker influence.

Changes Over Time

• Happiness scores showed regional stability but slight shifts in ranking.
• Some nations showed economic-driven increases/decreases in happiness.

Model Performance

• Our regression model achieved an R² score of 85%, indicating a strong predictive ability.

Author

Marouan El-Asery