How priyapaldatahub/weather-forecaster-ml works
This is a weather-temperature regression project: it explores and prepares historical climate data, trains and evaluates a Random Forest model, saves the model and feature mappings, and serves predictions through Streamlit. The main user path is weather inputs → prediction in the app → displayed temperature. Source excerpts show training invoking preprocessing and the app loading saved artifacts; README statements support the broader documented workflow. The README’s Parquet dataset is not present in the listed tree, so dataset links reflect documented inputs rather than a verified checked-in file.
Main entry points: app.py Input interface · train.py Train and evaluate · datacollection.py Dataset loader · preprocessing.py Clean and engineer
Components 11
Data exploration
- Dataset loader
src/datacollection.py - Weather analysis
src/EDA.py
Data preparation
- Clean and engineer
src/preprocessing.py
Model training
- Train and evaluate
src/train.py - Model artifacts
models/weather_model.pkl - Feature mappings
models/feature_columns.pkl
Prediction app
- Input interface
src/app.py - Temperature prediction
src/app.py
Other components
- Historical weather data
- Weather user
- Predicted temperature
Connections 12
- Historical weather data → Dataset loader: loads dataset
src/datacollection.py - Historical weather data → Weather analysis: analyzes data
src/EDA.py - Historical weather data → Clean and engineer: loads data
src/preprocessing.py - Dataset loader → Historical weather data: returns data
src/datacollection.py - Train and evaluate → Clean and engineer: calls cleaning
src/train.py - Train and evaluate → Model artifacts: saves model
src/train.py - Train and evaluate → Feature mappings: saves columns
src/train.py - Weather user → Input interface: enters weather
README.md - Input interface → Model artifacts: loads model
src/app.py - Input interface → Feature mappings: loads mappings
src/app.py - Input interface → Temperature prediction: submits inputs
src/app.py - Temperature prediction → Predicted temperature: displays temperature
src/app.py