Gender Inequality & Violence Analysis using Data Science

This project explores the relationship between gender inequality, societal attitudes, and violence against women using statistical analysis and machine learning techniques.

---

🚀 Project Overview

Violence against women is a global issue rooted in structural inequality and societal norms.

This project analyzes international datasets to:

• Understand the relationship between gender inequality and health outcomes
• Explore societal attitudes toward violence
• Identify patterns in women's education and empowerment

The goal is to provide data-driven insights into one of the most critical social challenges worldwide.

---

📂 Dataset

This project combines two datasets:

1. Gender Inequality Index (GII)

• 190+ countries
• Health, education, and economic indicators

2. Violence Against Women Dataset (DHS)

• 12,000+ survey responses
• Attitudes toward justification of violence
• Sociodemographic variables

Data was merged using country-level mapping and cleaned for consistency.

👉 Dataset sample:
View full dataset

---

🔬 Methodology

Data Wrangling

• Left join on country
• Standardization of country names
• Missing value imputation (median/mode)
• Duplicate removal

---

Exploratory Data Analysis

Gender Inequality vs Maternal Mortality

• Strong positive relationship between inequality and maternal mortality
• Countries with higher GII show significantly worse health outcomes

---

Violence Justification by Gender

• Women report higher justification levels
• Reflects internalized societal norms and cultural influence

---

Education vs Political Representation

• Weak but visible relationship between education and representation
• Indicates structural and societal factors beyond education

---

📊 Modeling

Linear Regression (OLS)

• Target: Maternal Mortality
• Predictor: Gender Inequality Index

Results:

• R² ≈ 52.1%
• +1 increase in GII → +1391 maternal deaths
• Strong statistical significance

---

Logistic Regression

• Predict gender based on violence justification

Results:

• Accuracy ≈ 59.8%
• Better performance for males than females
• Indicates limited predictive power of attitudes alone

---

🧠 Key Insights

• Gender inequality strongly impacts maternal health outcomes
• Societal norms influence acceptance of violence
• Education alone is not sufficient for empowerment
• Structural inequalities drive global disparities

---

🛠️ Tech Stack

---

🔮 Future Work

• Extend analysis to country-specific case studies (e.g., Türkiye)
• Improve predictive models with additional features
• Apply causal inference methods
• Incorporate time-series analysis

---

⚡ Implementation

All analysis and modeling code is available in the notebooks/Dashboard.ipynb directory.

---

👩‍💻 Author

Irem Akcan