Australian Weather Rain Prediction 🌧️

This project builds an end-to-end machine learning pipeline to predict whether it will rain tomorrow in Australia, using historical weather data. The project includes:

• Data cleaning and preprocessing
• Feature engineering
• Model training with scikit-learn
• Model evaluation
• An interactive Streamlit web application for:
  • Data inspection & cleaning preview
  • Real-time rain prediction

---

🚀 Features

• Modular project structure (no notebook dependency for training)
• Automated data cleaning
• Missing value handling using imputers
• Logistic Regression classification model
• Streamlit-based UI with:
  • Data Cleaning & EDA page
  • Rain Prediction page
• Reproducible training pipeline
• Production-ready model saving & loading

---

📂 Project Structure

WeatherForecast/
├── data/
│   └── raw/
│       └── AUS_Weather.csv
├── models/
│   └── weather_logreg.joblib
├── notebooks/
│   └── exploration.ipynb
├── notebooks/
│   ├── exploration.ipynb
│   └── exploration2.ipynb
├── src/
│   ├── __init__.py
│   ├── config.py
│   ├── data.py
│   ├── features.py
│   ├── train.py
│   └── predict.py
├── app.py
├── requirements.txt
├── README.md
└── .gitignore

🧠 Model Details

Algorithm: Logistic Regression

Target Variable: RainTomorrow

Selected Features:

MinTemp

MaxTemp

Humidity3pm

Pressure3pm

WindSpeed3pm

RainToday

Preprocessing:

Numerical features → Median imputation + StandardScaler

Categorical features → Mode imputation + OneHotEncoder

⚙️ How to Run

1️⃣ Install dependencies

pip install -r requirements.txt

2️⃣ Train the model

python -m src.train

This will generate:

models/weather_logreg.joblib

3️⃣ Run the Streamlit app

streamlit run app.py

🖥️ Application Pages

✅ Data Cleaning & EDA

Displays raw dataset

Shows missing values

• Install dependencies:

pip install -r requirements.txt

• Run the model script:

python src/exploration.ipynb

• Output

Displays cleaned dataset preview

✅ Prediction Page

User inputs weather conditions

Outputs:

Rain / No Rain

Prediction probability

📈 Future Enhancements

Batch predictions via CSV upload

Advanced models (XGBoost, Random Forest)

Model performance visualizations

Deployment on Streamlit Cloud

📜 License

This project is open-source and free to use for learning and experimentation.