Maternal Mortality Analysis (South West Nigeria)

This repository contains advanced data analysis and visualization code for exploring maternal mortality in South West Nigeria from 2000–2024.

Overview

The dataset captures hospital-level and regional maternal mortality records across teaching hospitals, private hospitals, Federal Medical Centres (FMCs), and specialty facilities. The analysis focuses on:

• Long-term mortality trends (2000–2024)
• Facility-level hotspots (top hospitals with most deaths)
• Geospatial analysis using Folium (mortality clusters and choropleths)
• Neonatal outcomes linked to maternal mortality
• Advanced statistical modeling (survival analysis, Cox proportional hazards, Kaplan-Meier)
• Predictive modeling with 7+ machine learning algorithms (Logistic Regression, Random Forest, SVM, Neural Nets, XGBoost, LightGBM, CatBoost)

Code Explanation

• trend_maternal_deaths.png → Line chart showing maternal deaths per year (2000–2024).
• hospital_hotspots.png → Bar chart showing the top 15 facilities with the highest maternal deaths.
• maternal_mortality_heatmap_stakeholders.html → Interactive heatmap highlighting mortality clusters across South West Nigeria.
• neonatal_outcome_trends.html → Area plot comparing neonatal survival vs deaths over time.
• Kaplan–Meier survival plots → Visualizing probability of survival after hospital admission, segmented by booking status.
• Cox Proportional Hazards Model → Statistical model identifying risk factors that increase hazard of maternal mortality.

Author

Deethepytor
Main contributor and code author.

Importance

This project is designed to help policymakers, healthcare stakeholders, and NGOs understand maternal mortality patterns and identify critical intervention points to reduce deaths and improve maternal care in Nigeria.

---