🏡 Boston House Price Predictor

---

A Streamlit-based web application that predicts median housing prices in Boston using Polynomial Regression, built from the Boston Housing Dataset. It supports interactive model tuning, residual visualization, and user-defined prediction.

https://boston-house-price-prediction-jnbhdqich2tsn3j9s2pgcf.streamlit.app/

📆 Project Overview

This project builds a machine learning web app that predicts housing prices using polynomial regression. The interface is built with Streamlit, and the model uses scikit-learn for preprocessing and training. The app features:

• Polynomial regression model
• Model evaluation metrics: MSE and R²
• Residual plot for model diagnostic
• Feature importance (top coefficients)
• Live prediction from user inputs
• Polynomial degree selector
• CI/CD-ready structure for deployment to Streamlit Cloud

---

📊 Dataset: Boston Housing Data

This dataset contains information collected by the U.S. Census Service concerning housing in the area of Boston, Massachusetts. It includes 506 instances and 14 attributes (13 features + target).

Target Variable:

• MEDV: Median value of owner-occupied homes in $1000s

Features Used:

Feature	Description
CRIM	Crime rate per capita
ZN	Proportion of residential land zoned
INDUS	Non-retail business acres per town
CHAS	Charles River dummy variable (1 if tract bounds river)
NOX	Nitric oxides concentration (parts per 10 million)
RM	Average number of rooms per dwelling
AGE	Proportion of owner-occupied units built prior to 1940
DIS	Weighted distances to employment centers
RAD	Index of accessibility to radial highways
TAX	Full-value property-tax rate per $10,000
PTRATIO	Pupil-teacher ratio by town
B	1000(Bk - 0.63)^2 (where Bk is % of Black residents)
LSTAT	% lower status population

---

🚀 Model and Features

Preprocessing

• StandardScaler is used to normalize all features to zero mean and unit variance.
• PolynomialFeatures is used to expand the feature space to include higher-order and interaction terms.

Model

• sklearn.linear_model.LinearRegression is used.

• Model is trained on a split of 80% training and 20% test data.

• Model evaluation uses:

  • Mean Squared Error (MSE)
  • R² Score (Variance Explained)

---

🌍 Web App UI (Streamlit)

Sidebar / Controls

• Slider to select polynomial degree (1 to 5)

• Expanders for:

  • Model evaluation
  • Residual plot
  • Feature importance (top 10 terms by coefficient)
  • User-defined prediction inputs

Output

• Real-time metrics update
• Matplotlib residual plot
• Top coefficients table
• House price prediction from user input

---

🛋️ Files in the Project

.
├── app.py                  # Streamlit frontend UI
├── model.py                # Model training and return function
├── requirements.txt        # Python dependencies
├── Boston-house-price-data.csv # Dataset file (13 features + MEDV)
└── .github
    └── workflows
        └── deploy.yml      # GitHub Actions for Streamlit Cloud deploy

---

🎓 How to Run Locally

1. Install Python 3.10+

2. Clone the repository:

   git clone https://github.com/your-username/boston-house-price-predictor.git
   cd boston-house-price-predictor

3. Install dependencies:

   pip install -r requirements.txt

4. Run Streamlit app:

   streamlit run app.py

---

🚫 Limitations

• Assumes feature relationships are polynomial in nature.
• Does not include regularization (e.g., Ridge/Lasso).
• Dataset is static and not updated with real-world values.

---

🤝 Credits

• Dataset: UCI Machine Learning Repository
• UI: Streamlit
• ML: scikit-learn

---

📊 Future Improvements

• Add Ridge/Lasso models
• Include SHAP explainability
• Add dataset uploader for user data
• Export predictions as CSV

---

Made with ❤️ by Mansoob E Zehra