🏠 Housing Price Prediction

This is my very first machine learning project, where I predict house prices based on input features using supervised regression models. I followed a project-based and self-learning approach to study AI. Although I'm not yet well-versed in ML theory, I aimed to create a complete and functional end-to-end workflow. ❤️

---

🧠 Problem Definition

The goal of this project is to:

• Predict housing prices based on features (area, bedrooms, parking, etc.).
• Apply a real-world machine learning pipeline.
• (Optionally in the future) infer which features most impact housing prices — useful for real estate and business decisions.

---

📊 Dataset

• Source: Kaggle
• Format: .zip containing a .csv file
• Features:
  • Numerical: area, bedrooms, bathrooms, stories, parking, price
  • Categorical (binary): mainroad, guestroom, basement, hotwaterheating, airconditioning, prefarea
  • Categorical (nominal): furnishingstatus (furnished/semi/unfurnished)

---

🔁 Project Workflow

1. Data Collection & Reading
2. Data Cleaning
   • Removed 15 outliers using IQR
   • No missing or duplicate values
3. Data Preprocessing
   • One-hot encoding for categorical features
   • Feature scaling with StandardScaler
4. Exploratory Data Analysis (EDA)
   • Distribution of price
   • Correlation and boxplots
   • Multicollinearity check
5. Feature Engineering
   • Created sum_area = area × stories
   • Created amenities_count = count of all "yes" binary features
   • One-hot encoded stories
6. Modeling
   • Baseline: LinearRegression
   • Final: XGBoostRegressor
7. Evaluation & Tuning
   • Used RMSE, R²
   • GridSearchCV and RandomSearchCV for tuning
   • Visualized predicted vs actual values

---

✅ Final Results

• Model: XGBoostRegressor
• RMSE: 1,093,408
• R² Score: 0.653

---

📈 Actual vs Predicted Plot

---

🛠 Technologies Used

• Python
• Pandas, NumPy, Seaborn, Matplotlib
• Scikit-learn
• XGBoost
• Google Colab

---

🧑‍💻 How to Run

# 1. Clone the repository
git clone https://github.com/your-username/Housing-price-prediction.git

# 2. Upload the zip dataset from Kaggle

# 3. Open and run the notebook in Google Colab or Jupyter