Weather Prediction Model – Linear Regression

Overview

This project is a machine learning weather prediction system built using Python, Scikit-Learn, and Streamlit.
The model predicts the temperature (°C) using weather-related features such as humidity, wind speed, visibility, pressure, and precipitation type.

The pipeline includes:

• Data preprocessing
• Feature engineering
• Linear Regression model training
• Temperature prediction script
• Interactive Streamlit Web App

---

Key Features

• Temperature Prediction using Linear Regression
• Preprocessing Pipeline (scaling, encoding, handling missing values)
• Time Feature Extraction (year, month, day, hour, day of week)
• Model Evaluation (MAE, RMSE, R²)
• Streamlit Web App for real-time predictions
• Modular Codebase (clean structure for training & inference)

---

Tech Stack

• Language: Python
• ML Framework: Scikit-Learn
• Web App: Streamlit
• Data Handling: pandas, NumPy
• Model Saving: joblib

---

Project Structure

Weather_Prediction/
│
├── data/
│   └── weatherHistory.csv          
│
├── models/
│   └── weather_model.joblib        
│
├── src/
│   ├── __init__.py
│   ├── preprocess.py                
│   ├── train_model.py              
│   ├── predict_example.py          
│   └── app.py                      
│
├── requirements.txt                
└── README.md

---

Model Performance

After training on the provided dataset:

• MAE: ~0.74°C
• RMSE: ~0.95°C
• R² Score: ~0.99

---

How the Model Works

Input Features

• Apparent Temperature
• Humidity
• Wind Speed
• Wind Bearing
• Visibility
• Pressure
• Precipitation Type
• Derived Date-Time Features (year, month, day, hour, dayofweek)

Output

• Predicted Temperature (°C)

---

Installation

1. Clone the repository:

git clone https://github.com/DulinaSamarathunga/Weather-Prediction-Model.git
cd Weather-Prediction-Model

2. Install dependencies:

pip install -r requirements.txt

---

Train the Model

python src/train_model.py

This trains the Linear Regression model and saves it to:

models/weather_model.joblib

---

Make a Prediction (Script)

python src/predict_example.py

Example output:

Predicted Temperature (C): 22.65

---

Run the Streamlit Web App

streamlit run src/app.py

App will open at:

http://localhost:8501

---

Requirements

• Python 3.12+
• pandas
• numpy
• scikit-learn
• joblib
• streamlit

---

Future Improvements

• Add Random Forest or XGBoost models
• Deploy Streamlit app online
• Improve UI with charts
• Add a REST API endpoint

---

Author

Dulina Samarathunga
GitHub: https://github.com/DulinaSamarathunga