How zahinsyukran/dengue-breeding-prediction-fyp works
This is a Taiwan dengue early-warning dashboard: users select a district, date, and model; the app obtains weather inputs, estimates risk, and presents maps, summaries, and a forecast calendar. The README documents a separate spatially agnostic training workflow and its validation approach. Source excerpts are partial, so the graph omits uncertain internal wiring and treats documented training stages at notebook level.
Main entry points: app.py Streamlit dashboard · model_train.ipynb Model training · models XGBoost assets · taiwan_dengue_master_dataset.csv Taiwan master data
Components 17
Data and training
- Taiwan master data
backend/data/taiwan_dengue_master_dataset.csv - Model training
backend/model_train.ipynb - Temporal aggregation
backend/model_train.ipynb - Robust validation
backend/model_train.ipynb
Forecast runtime
- Streamlit dashboard
frontend/app.py - OpenWeather API
- NASA POWER API
- Firebase config
- Feature engineering
frontend/app.py - Risk prediction
frontend/app.py
Model assets
- XGBoost assets
backend/models - LSTM assets
backend/models
Dashboard results
- Risk map
frontend/app.py - Risk summaries
frontend/app.py - Weather overview
frontend/app.py - Forecast calendar
frontend/app.py
Other components
- Dashboard user
Connections 19
- Taiwan master data → Model training: trains from
- Model training → Temporal aggregation: aggregates by date
- Model training → Robust validation: validates models
- Model training → XGBoost assets: saves model assets
- Model training → LSTM assets: saves model assets
- Dashboard user → Streamlit dashboard: selects options
- Streamlit dashboard → OpenWeather API: requests forecast
- Streamlit dashboard → NASA POWER API: requests history
- Streamlit dashboard → Firebase config: reads NDVI config
- Streamlit dashboard → XGBoost assets: loads assets
- Streamlit dashboard → LSTM assets: loads assets
- Streamlit dashboard → Feature engineering: engineers features
- Feature engineering → Risk prediction: provides features
- Risk prediction → Risk map: provides risk levels
- Risk prediction → Risk summaries: provides risk results
- Streamlit dashboard → Weather overview: displays weather
- Streamlit dashboard → Forecast calendar: builds calendar
- Dashboard user → Risk map: views forecast
- Dashboard user → Risk summaries: views summary