ekpenyongasuquo/nigeria-maternal-risk-oracle

Predicting maternal health risk in Nigeria using WHO/DHS indicators — ZerveHack 2026 # 🇳🇬 Nigeria Maternal Risk Oracle ### ZerveHack 2026 Submission | $10,000 Prize Competition ## The Question 52% of Nigerian women complete antenatal care. Only 43% deliver in a health facility. Who is most at risk — and what should policymakers do first? ## Live Demo 🔗 **Interactive App:** ng-risk-oracle.hub.zerve.cl… 🔗 **Zerve Project:** app.zerve.ai ## The ANC Paradox Nigerian women attend checkups faithfully — then go home to deliver alone. This structural dropout between antenatal care and facility delivery is the hidden driver of Nigeria's maternal mortality crisis. This oracle quantifies that gap and predicts risk in real time. ## Key Findings (2000–2021) | Finding | Value | |---|---| | ANC-to-Facility Gap (2000) | 14.8 percentage points | | ANC-to-Facility Gap (2021) | 7.0 percentage points | | MMR Improvement | 45% decline (1,210 → 659 per 100k) | | Health Expenditure vs MMR | r = −0.78 (strong correlation) | | Top Policy Lever | Skilled birth attendance +10pp | ## How It Works Input any population's health indicators → get instant: - ✅ Risk classification (AT RISK / SAFE) - ✅ ANC-to-facility gap measurement - ✅ Top contributing risk factor - ✅ Prioritised policy recommendation ## Data Sources | Source | Institution | Coverage | |---|---|---| | Maternal & Reproductive Health Indicators | WHO / HDX | 2000–2021 | | Health Financing Database | WHO / HDX | 2000–2023 | | Maternal Mortality Survey | DHS Nigeria | 2008–2018 | ## Try It Live Visit the app and move the sliders to simulate different health system scenarios: 🔗 ng-risk-oracle.hub.zerve.cl… ## Tech Stack - **Analysis:** Python, Pandas, NumPy, Scikit-learn - **Platform:** Zerve AI (autonomous reasoning + deployment) - **App:** Streamlit - **Data:** WHO HDX, DHS Program ## Author Built for ZerveHack 2026 — Zerve AI Hackathon Deadline: 29 April 2026 Prize Pool: $10,000