Nyuira/HIV-PMTCT-Optimization

Forecasting HIV Positivity Rates and Optimizing PMTCT Resource Allocation in Kenya Here's the **complete README.md** content - copy and paste this entire block: ```markdown # HIV Positivity Rate Forecasting & PMTCT Resource Optimization in Kenya ## Project Overview This project aims to forecast HIV positivity rates in Maternal and Child Health (MCH) services across Kenyan facilities and optimize resource allocation for Prevention of Mother-to-Child Transmission (PMTCT) interventions. Using historical MOH 731 data (2020-2024), I predict trends and provide actionable insights to help Kenya achieve its <5% MTCT target by 2030. ## Objectives 1. **Predict** HIV positivity rates with MAE ≤ 2% using XGBoost and Prophet 2. **Analyze** trends in MCH HIV metrics to identify high-risk facilities 3. **Optimize** resource allocation (ART, testing kits) to reduce projected positives by ≥20% 4. **Deploy** interactive dashboard (Streamlit) and API (FastAPI) for stakeholder use ## Dataset - **Source**: MOH 731 reports (NSDCC, KHIS) - **Files**: - `mch proportion.csv`: 3,553 facility-level records (8 features) - `HIV_dataset.csv`: 1,828 aggregated records (40 features) - Additional datasets: [To be added as project progresses] - **Time Period**: 2020-2024 - **Geographic Coverage**: Kenyan counties and facilities ## Project Structure ``` HIV-PMTCT-Optimization/ ├── data/ # Data directories │ ├── raw/ # Original CSV files │ ├── processed/ # Cleaned and merged data │ └── synthetic/ # Synthetic data for reproducibility ├── notebooks/ # Jupyter notebooks for each phase │ ├── 01_data_preprocessing.ipynb │ ├── 02_exploratory_data_analysis.ipynb │ ├── 03_model_development.ipynb │ ├── 04_optimization.ipynb │ └── 05_evaluation.ipynb ├── src/ # Source code modules │ ├── data_preprocessing.py │ ├── feature_engineering.py │ ├── models.py │ ├── optimization.py │ └── utils.py ├── deployment/ # Deployment files │ ├── streamlit_app.py # Inte …