CAR_PRODUCT_ANALYSIS

🚗 Car Product Analysis

A full-stack analytics solution designed to uncover pricing dynamics, mileage trends, and brand valuation in the automobile resale market. This project empowers dealerships, resale platforms, and analysts with predictive modeling and interactive dashboards for smarter pricing and inventory decisions.

---

🚗 GitHub Project Repository

🔗 Click to view CAR_PRODUCT_ANALYSIS

---

🧠 Project Overview

The resale value of a car depends on multiple factors—brand, mileage, condition, and age. This project analyzes 100,000+ car listings to build a regression model that predicts price and visualizes key market insights.

Key Objectives:

• Clean and preprocess car listing data
• Engineer features for price modeling and dashboarding
• Build regression models to predict car resale value
• Deploy interactive dashboards for business decision-making

---

📁 Project Structure

File Name	Description
Car_Data.csv	Raw dataset containing 100,000 car listings
cleaned_car_sales.csv	Preprocessed dataset used for modeling
car product analysis.ipynb	Jupyter notebook with full analysis workflow
car_product_analysis.sql	SQL queries for data extraction and transformation
sqlconnect.py	Python script for SQL database connection
car_price_model.pkl	Trained machine learning model for price prediction
feature_names.pkl	Serialized feature list used in model training
label_encoders.pkl	Encoders for categorical variables
app.py	Streamlit app for interactive model deployment
Car Product Analysis DASHBOARD.twb	Tableau dashboard visualizing key metrics and trends

---

🧹 Data Preprocessing

• Imputed missing values in Mileage and Condition
• Created derived features like Car Age from Year
• Encoded categorical variables (Brand, Model, Color, Condition)
• Removed outliers in Price and Mileage
• Optimized data types for memory efficiency

---

📈 Exploratory Data Analysis

• Price distribution across brands and years
• Mileage trends segmented by car condition
• Age-based depreciation curves
• Condition impact on pricing
• Correlation matrix of features influencing price

---

🤖 Modeling Approach

• Target Variable: Price
• Algorithms Used: Linear Regression, Random Forest, XGBoost
• Evaluation Metrics: RMSE, MAE, R² Score
• Feature Importance: Car Age, Mileage, Brand, Condition, Model

---

📊 Dashboard Overview

📍 Tableau Dashboard

Visualizes key pricing and mileage metrics:

• 📊 Price distribution by brand and year
• 🚗 Mileage trends segmented by condition
• 📈 KPI cards for average price, mileage, and car age
• 🧠 Feature impact on predicted price

---

🟢 Streamlit App

Interactive model deployment with real-time predictions:

• 🛠️ Input car attributes to get predicted price
• 📉 Visual breakdown of feature importance
• 🔍 Dynamic filtering by brand, condition, and year
• 🧮 SQL-integrated querying for custom car segments

---

🚀 Deployment

• Model serialized with joblib as car_price_model.pkl
• Dashboard deployed via Streamlit Cloud
• Tableau dashboard built using .twb workbook
• SQL integration for dynamic data querying
• Git LFS used for large file management

---

🧠 Business Impact

• Enables accurate pricing for resale platforms and dealerships
• Identifies undervalued or overvalued listings
• Supports inventory planning based on brand and condition trends
• Enhances customer trust with data-driven pricing transparency

---

🛠️ Tech Stack

• Python: Pandas, NumPy, Scikit-learn, Streamlit
• SQL: Data extraction and transformation
• Visualization: Tableau, Matplotlib, Seaborn
• Deployment: Streamlit Cloud, GitHub, Git LFS

---

📌 Future Enhancements

• Integrate web scraping for real-time car listings
• Add classification model for condition prediction
• Enable user-uploaded car specs for instant valuation
• Expand dashboard to include brand-wise ROI and depreciation curves

---

👤 Author

Anesh Raj
Data Analyst | Data Scientist | Business Analyst
Focused on multi-industry impact through predictive modeling and dashboarding.
📍 Chennai, India
🔗 GitHub Profile