aminahol/africa-mortality-clustering

This project analyzes healthcare disparities in African countries using EDA, PCA, and KMeans clustering. It groups countries by population, GDP, mortality, and access to doctors, revealing outliers and patterns to support data driven policy and global health planning. # Clustering African Countries by Healthcare Access and Outcomes ## Project Overview This project analyzes healthcare disparities across African countries using data driven techniques. By integrating six datasets including population, GDP, access to doctors, causes of mortality, and mortality by age group it applies Exploratory Data Analysis (EDA), Principal Component Analysis (PCA), and KMeans Clustering to group countries with similar healthcare characteristics and identify outliers with critical health concerns. --- ## Objectives - Group African countries by healthcare and socio economic patterns - Identify outlier nations with extreme conditions - Reveal gaps in healthcare access across the continent - Support targeted interventions and data driven global health strategies --- ## Use Cases - Guide targeted health interventions in vulnerable regions - Inform strategic investments and resource allocation - Address healthcare inequalities at scale - Enable AI driven and data-informed public health solutions --- ## Tools and Libraries - **Python**: `pandas`, `numpy`, `scikit-learn`, `matplotlib`, `seaborn`, `plotly`, `kneed` - **Power BI**: for interactive, country-level and cluster-level dashboards --- ## Methodology 1. **Data Collection** Merged 6 datasets on GDP, population, mortality, causes of death, mortality by age group, and access to medical professionals. 2. **Data Cleaning and Preparation** Addressed missing values and inconsistencies, standardized country names, converted types, and scaled data for analysis. 3. **Exploratory Data Analysis (EDA)** Explored patterns and anomalies in mortality, population, GDP, and healthcare coverage. 4. **Dimensionality Reduction** Applied PCA, preserving 96.1% of total variance, to simplify the dataset and enhance clustering accuracy. 5. **Clustering** Performed KMeans Clustering to categorize countries into four distinct clusters based on healthcare indicators. 6. **Evaluation** - Silhouette Score: `0.82` …