🏠 House Sales Analysis in King County, USA- EDA + Regression Project

📑 Table of Contents

• Project Overview
• Objectives
• Dataset
• Features
• Tools & Libraries
• Exploratory Data Analysis(EDA)
• Model Evaluation & Refinement
• Key Insights
• Project Structure
• How to Run the Project
• Future Improvements
• Author & Contact

---

📌 Project Overview

This project involves:

• Performing Exploratory Data Analysis (EDA) to identify trends, patterns, and correlations.

• Handling missing data and preparing the dataset.

• Building and evaluating regression models (Linear Regression, Polynomial Regression, Ridge Regression).

• Comparing model performance to find the best approach for predicting house prices.

---

🎯 Objectives

The objective of this project is to analyze residential housing data from King County, USA (including Seattle) and predict house sale prices using regression models. The project simulates a real-world scenario where a Real Estate Investment Trust wants to estimate property values based on features like square footage, number of bedrooms, bathrooms, grade, and location.

---

🗂️ Dataset

• Source: King County Housing dataset (modified for learning purposes)

• Rows: 21,613

• Columns: 22

• Target Variable: price

--

Features:

• bedrooms, bathrooms

• sqft_living, sqft_lot

• floors, waterfront, view, condition, grade

• sqft_above, sqft_basement

• yr_built, yr_renovated, zipcode, lat, long

• sqft_living15, sqft_lot15

---

🔧 Tools & Libraries

• Python 🐍
• Pandas – Data manipulation & analysis
• Matplotlib,Seaborn – Data visualization
• Scikit-learn (Linear Regression, Ridge Regression, Polynomial Features, Pipelines) - Model Development
• Jupyter Notebook – Development & exploration
• Github

---

🔎 Exploratory Data Analysis (EDA)

• Checked missing values (bedrooms, bathrooms) and replaced with mean.

• Dropped irrelevant columns (id, Unnamed: 0).

• Distribution of houses by number of floors.

• Boxplot: waterfront vs price → waterfront houses tend to be more expensive.

• Regression plot: sqft_above vs price → strong positive correlation.

• Correlation heatmap → sqft_living, grade, and sqft_above most correlated with price.

---

📈 Model Development

1. Simple Linear Regression

• Feature: long → very weak predictor (R² ≈ 0.00046).

• Feature: sqft_living → moderate predictor (R² ≈ 0.49).

2. Multiple Linear Regression

• Features: floors, waterfront, lat, bedrooms, sqft_basement, view, bathrooms, sqft_living15, sqft_above, grade, sqft_living.

• R² ≈ 0.65

3. Polynomial Regression with Pipeline

• Degree = 2

• R² ≈ 0.75

---

Model Evaluation & Refinement

• Train-Test Split (85% train / 15% test).

• Ridge Regression (α=0.1): R² ≈ 0.64 on test data.

• Polynomial Features + Ridge Regression: R² ≈ 0.70 on test data.

--

📌 Key Insights

• Most important predictors:sqft_living, grade, sqft_above, sqft_living15.

• Waterfront houses and homes with higher grade have significantly higher prices.

• Polynomial regression improves performance compared to simple linear models.

📂 Project Structure

📦 House-Sale-Analysis-EDA-Regression-Python
│
├── README.md
├── .gitignore
├── notebooks/                  # Jupyter notebooks
│   ├── regression_exploratory_data_analysis.ipynb
├── data/                
    └──kc_house_data_NaN.csv

---

🚀 How to Run the Project

1. Clone the repository:

   git clone https://github.com/codewithchirag18/House-Sale-Analysis-EDA-Regression-Python.git

2. Navigate to the folder:

bash
Copy code
cd sales-analysis-eda

3. Install required libraries:

bash
Copy code
pip install -r requirements.txt

4. Open Jupyter Notebook:

bash
Copy code
jupyter notebook regression-exploratory_data_analysis.ipynb

---

📌 Future Work

• Try Random Forest or Gradient Boosting for better accuracy.

• Feature engineering (e.g., combine year built + renovation into “house age”).

• Deploy the model using Flask / Streamlit for predictions.

---

Author & Contact

Chirag Tomar

Data Analyst

📧 Email: tomarchirag431@gmail.com

🔗 LinkedIn

🔗 LeetCode

--