🏡 Housing Price Prediction

This project explores a housing dataset and builds a machine learning model to predict median house values based on demographic and geographic features.

The focus is on combining exploratory data analysis (EDA) with a clean modeling pipeline to understand what drives housing prices and establish a strong baseline model.

---

🚀 Project Overview

This notebook walks through an end-to-end workflow:

• Data loading and exploration
• Visualization of feature distributions and relationships
• Handling missing data
• Encoding categorical variables
• Building a preprocessing + modeling pipeline
• Training and evaluating a regression model

The final output is a Linear Regression model evaluated using RMSE.

---

📊 Key Insights

Some high-level takeaways from the analysis:

• Median income shows a strong positive relationship with house value
• Location-based features (latitude/longitude, ocean proximity) play a significant role
• Several features are right-skewed, suggesting potential for transformation in future iterations

---

📁 Project Structure

housing-price-prediction/
├── housing.csv                  # Dataset
├── housing_model.ipynb          # Main notebook (EDA + modeling)
├── README.md                    # Project documentation

---

🧠 Modeling Approach

• Model: Linear Regression

• Preprocessing:

  • Missing value imputation (median)
  • One-hot encoding for categorical features

• Pipeline: ColumnTransformer + Pipeline (scikit-learn)

• Train/Test Split: 80/20

---

📈 Evaluation

Model performance is evaluated using:

• Root Mean Squared Error (RMSE)

This provides an interpretable measure of average prediction error in housing price units.

---

🛠️ Tech Stack

• Python
• pandas, numpy
• matplotlib, seaborn
• scikit-learn

---

▶️ How to Run

1. Clone the repository:

git clone https://github.com/pmcelroy4/house_price_model.git
cd housing-price-prediction

2. Install dependencies (if applicable):

pip install -r requirements.txt

3. Launch the notebook:

jupyter notebook housing_model.ipynb

---

🔮 Future Improvements

• Try more advanced models (Random Forest, Gradient Boosting)
• Add feature engineering (e.g., rooms per household, bedrooms per room)
• Perform cross-validation for more robust evaluation
• Apply scaling or transformations to skewed features

---

💡 Why This Project

This project demonstrates:

• A structured approach to EDA and feature understanding
• Practical use of scikit-learn pipelines
• The ability to move from raw data → insights → model