🏠 House Price Prediction Web App

This project is a machine learning-based web application built with Streamlit for predicting house prices. It uses Linear Regression and optionally XGBoost, trained on housing data with preprocessing steps included in a pipeline.

---

📁 Project Structure

House\_Price\_Prediction/
│
├── dataset/
│   ├── train.csv
│   ├── test.csv
│   └── sample\_submission.csv
│
├── oneshot\_linear.py             # Train linear regression model
│
├── dataset/
│   ├── train.csv
│   ├── test.csv
│   └── sample\_submission.csv
│
├── train\_xgboost.py             # Train XGBoost model (optional)
├── app.py                       # Streamlit frontend app
├── zero\_shot\_model.pkl          # Trained model file (e.g., Linear or XGBoost)
│
├── dataset/
│   ├── train.csv
│   ├── test.csv
│   └── sample\_submission.csv
│
├── submission\_oneshot\_linear.csv # Generated submission file
├── requirements.txt
└── README.md

---

🚀 Features

• Upload custom house data and get predicted price
• Interactive sidebar for input features
• Visualizations:
  • Histogram of Sale Prices
  • Correlation Heatmap
  • Feature Importance (for XGBoost models)
• Preprocessing includes:
  • Missing value handling
  • One-hot encoding
  • Model serialization using joblib

---

📦 Requirements

Install dependencies using:

pip install -r requirements.txt

requirements.txt

pandas
scikit-learn
joblib
xgboost
streamlit
matplotlib
seaborn

---

🧠 Training the Model

➤ Linear Regression

python oneshot_linear.py

This will:

• Train a model on train.csv
• Save predictions to submission_oneshot_linear.csv
• Export trained model as oneshot_linear_model.pkl

➤ XGBoost (optional)

python train_xgboost.py

Exports model as xgboost_model.pkl

---

💻 Running the Web App

To launch the Streamlit frontend:

streamlit run app.py

Make sure your trained model (e.g., zero_shot_model.pkl) is present in the directory.

---

🔍 Example Inputs

• Numerical Inputs: LotArea, GrLivArea, YearBuilt, etc.
• Categorical Inputs: Neighborhood, HouseStyle, etc.

---

📊 Screenshots

---

🤖 Model Notes

• Linear Regression gives a good baseline performance.
• XGBoost supports feature importance plotting.
• Trained pipeline includes full preprocessing using ColumnTransformer.

---

🧾 License

This project is licensed under the MIT License.

---

🙋‍♂️ Author

Jayasimma D

Feel free to connect on LinkedIn or contribute to this project!