How kaiser287/cyclone-intensity works
This repository supports cyclone wind-intensity prediction, with a Streamlit dashboard that accepts satellite imagery, runs inference, and presents intensity and track information. A separate training workflow prepares image data, builds an intensity model, and trains it; the repository also includes HDF5 preprocessing and an automatic Excel reporting script. The README documents manual TCIR data setup and training. Some data and training modules were not excerpted, so their internal connections are omitted where the evidence does not establish them.
Main entry points: main.py Training Entry · app.py Prediction Dashboard · dataset.py Image Dataset · predictor.py Intensity Predictor
Components 15
Dashboard and Reports
- Prediction Dashboard
app.py - Metrics and Maps
app.py - Excel Report
generate_report_auto.py
Data Preparation
- HDF5 Preprocessing
source/data/preprocessor.py - Image Dataset
source/training/dataset.py
Model Training
- Training Settings
config/settings.py - Training Entry
main.py - Training Loop
source/training/trainer.py - Intensity Training
source/training/train_intensity.py - Track Training
source/training/train_track.py
Inference Models
- Intensity Model
source/models/intensity_model.py - Intensity Predictor
source/inference/predictor.py - Track Predictor
source/models/track_model.py
Other components
- Analyst
- Satellite Data
Connections 15
- Analyst → Prediction Dashboard: uploads imagery
app.py - Satellite Data → HDF5 Preprocessing: provides HDF5
README.md - Prediction Dashboard → Intensity Predictor: runs inference
app.py - Intensity Predictor → Track Predictor: initializes track
source/inference/predictor.py - Prediction Dashboard → Metrics and Maps: renders results
app.py - Excel Report → Intensity Predictor: requests predictions
generate_report_auto.py - Satellite Data → Image Dataset: provides images
README.md - Training Entry → Training Settings: reads config
main.py - Training Entry → Image Dataset: creates dataset
main.py - Training Entry → HDF5 Preprocessing: sets transforms
main.py - Image Dataset → HDF5 Preprocessing: uses transforms
source/training/dataset.py - Training Entry → Intensity Model: creates model
main.py - Training Entry → Training Loop: starts training
main.py - Training Loop → Training Settings: reads config
source/training/trainer.py - Track Training → Track Predictor: uses model
source/training/train_track.py