Built as part of the AI4Good Lab – Toronto Cohort (2025), Team 3 Project
AI-powered climate intelligence for global health Predict tomorrow’s Neglected Tropical Disease (NTD) hotspots before outbreaks happen.
🚀 Explore the Deployed Website
Climate change is accelerating the emergence and migration of diseases - reshaping boundaries faster than traditional public health systems can adapt.
Zoniq integrates:
- 🌡️ Real-time climate data
- 🧬 Epidemiological patterns
- 🤖 Machine learning forecasting
Together, these power a platform that anticipates NTD risks before outbreaks, enabling timely interventions and data-driven public health strategy.
Our live platform showcases:
- 📍 Visualize predicted NTD hotspots in Nigeria (2026–2028)
- 📈 XGBoost-driven forecasts grounded in climate and disease data
- 🌎 Expansion plans to model disease risk worldwide
Cleaned and merged dataset containing:
-
Town-level health, population, and climate attributes
-
Used as the input foundation for all model training and visualization
Summary of:
-
Data cleaning and transformation
-
Feature engineering methodology
-
Justification for variable selection and modeling decisions
Final presentation deck delivered for:
🧠 AI4Good Lab – Toronto Demo Day
💡 Project background, motivation, methodology, and impact vision
A detailed Jupyter Notebook containing:
-
Benchmarked models:
-
Logistic Regression
-
Random Forest
-
XGBoost
-
PCA
-
Performance metrics (precision, recall, F1)
-
Data inputs and generated reports
| Model | Precision (Class 1) | Recall (Class 1) | F1 Score (Class 1) | Accuracy |
|---|---|---|---|---|
| Logistic Regression | 0.71 | 0.11 | 0.20 | 81.5% |
| Random Forest | 0.53 | 0.44 | 0.48 | 81.2% |
| XGBoost (Best) | 0.67 | 0.72 | 0.70 | 87.0% |
- WHO NTD surveillance data
- ERA5 climate reanalysis datasets
- Local health and demographic surveys
- Custom engineered features from spatial-temporal trends
MIT License © 2025 — Zoniq AI Team