toffickm252/sage-hypertension-predictor-ssa

Predicting hypertension risk in Sub-Saharan African adults using WHO SAGE data from Ghana and South Africa # sage-hypertension-predictor-ssa Predicting hypertension risk in Sub-Saharan African adults using WHO SAGE data from Ghana and South Africa ## Overview Most hypertension prediction models are trained on Western or Asian populations. This project builds a localized, evidence-based risk tool using real measured health data from Ghanaian and South African adults. Built as part of the ENg 30-Day Build in Public Challenge (Data/AI Track). ## Dataset WHO Study on Global Ageing and Adult Health (SAGE) — Wave 1 (2007–2010) Countries: Ghana and South Africa Source: WHO Study on Global AGEing… Blood pressure classification follows the 2018 ESC/ESH staging system as adopted in Ghana's national cardiovascular disease guidelines. ## Pipeline 1. Data ingestion and loading 2. Preprocessing — missing values, outliers, data types 3. Feature engineering — age-sex interaction terms, BP staging, BMI categories 4. Exploratory data analysis — hypertension trends across age, sex, and country 5. Model training — logistic regression (baseline), random forest, XGBoost 6. Evaluation and comparison — accuracy, precision, recall, AUC-ROC 7. Streamlit web app — user inputs lifestyle data, receives risk classification ## Deliverables - Deployed Streamlit web app - Documented Jupyter notebook - Written summary of findings ## Progress Built in public. Daily updates on X: @elzer252 #ENg30DayChallenge #ENgShipIt ## Setup ## Instructions will be added as the project starts.