Used Car Price Prediction

Created2025-07-24|Updated2025-08-12|Showcase

|Post Views:

🚗 Used Car Price Prediction in Virginia

Predicting the price of used cars with AI and data analysis

“Don’t guess the price — let the data tell you.”

📎 Full Analysis:
👉 View Jupyter Notebook on GitHub

Web App

📌 One-Line Summary

This project predicts the prices of used cars in Virginia using a dataset of over 46,000 listings.
By analyzing details like year, mileage, brand, and fuel type, the AI model can estimate a realistic market price.

1️⃣ How It Was Built

1. Data Collection

Collected real car sales data from the web
Stored in an AWS cloud MySQL database
Accessed using Python and shared via Flask API

2. Data Preparation

Filled missing values (e.g., unknown mileage)
Removed unrealistic values (e.g., mileage over 1 million km)
Converted text data (brand, fuel type) into numbers
Applied log transformation to balance skewed data

3. Data Analysis (EDA)

Visualized the relationship between price and year/mileage
Found year and mileage to be the most influential features

4. AI Model Training

Tested several machine learning models:

Linear Regression
Decision Tree
Random Forest
Support Vector Regression (SVR)
XGBoost (winner)

📊 Best Model: XGBoost

Accuracy (R²): 0.89
Average Error (RMSE): $5,474

5. Results

Model	R² Score	RMSE
Linear Regression	0.58	14,085
Decision Tree	0.73	9,022
Random Forest	0.84	7,021
SVR	0.11	14,328
XGBoost	0.89	5,474

6. Real-World Test

2016 Honda Odyssey → Predicted price: $18,738
Matched closely with actual market data.

2️⃣ Real-World Use

Final model saved as a Pickle file
Deployed via Flask API for real-time predictions
Created a simplified version (Year, Mileage, Brand, Model) for web app integration

🛠 Technologies Used

Step	Technology
Data Storage	AWS MySQL
Model Dev	Python, scikit-learn, XGBoost
Deployment	Flask API, Pickle
Environment	AWS EC2 (Ubuntu)

💡 Key Learnings

Log transformation improves accuracy for skewed data
Tree-based models handle mixed data types effectively
Even with only 4 features, accurate real-time predictions are possible

🔗 GitHub Repository

📂 View Project on GitHub

Author: Danny Ki

Link: https://kish191919.github.io/2025/07/24/Used-Car-Price-Prediction-in-Virginia/

Copyright Notice: All articles on this blog are licensed under CC BY-NC-SA 4.0 unless otherwise stated.

Python Machine Learning XGBoost

Related Articles

House Price Prediction

🏠 House Price Prediction in Ames, IowaPredicting real estate prices using advanced regression techniques “Accurate home valuation — powered by data.” 📎 Full Analysis:👉 View Jupyter Notebook on GitHub 📎 Competition Source:👉 Kaggle: House Prices - Advanced Regression Techniques 📌 One-Line SummaryThis project predicts house prices in Ames, Iowa using 81 property features such as area, location, condition, and building type.It applies data preprocessing, feature engineering, and re...

[Algorithm] Bubble Sort — Can Life Be Sorted Too?

🫧 What’s the Big Deal About Bubble Sort? “It’s the little differences that make the big difference.” – Napoleon Hill In the programming world, there are moments when small changes lead to big outcomes—just as Napoleon Hill said.One of the best examples of this is Bubble Sort. At first glance, it’s so simple that it’s easy to dismiss. But behind its simplicity lies a surprisingly meaningful lesson. Bubble Sort works by scanning a list from start to finish, comparing two neighboring e...

[Algorithm] Dynamic Programming — Trusting Your Past Self

🔁 What’s the Big Deal About Dynamic Programming? “Memory is the treasury and guardian of all things.” – William Shakespeare In life, there are moments when we need to remember and refer to past experiencesto avoid making the same mistakes again. Whether in studying, exercising, or love—the ability to remember and use what we’ve learned shapes our future. There’s an algorithm in programming that reflects this very philosophy:Dynamic Programming (DP). 🧠 The Philosophy Behind ItDyn...

[Algorithm] Insertion Sort — Quiet but Steady Growth

🧩 What’s the Big Deal About Insertion Sort? “It does not matter how slowly you go as long as you do not stop.” – Confucius In life, there are those who don’t seek the spotlight but quietly find their own place.In the world of algorithms, there’s a counterpart to such people — Insertion Sort. As its name suggests, Insertion Sort works by inserting each new elementinto the correct position within an already sorted portion of the list. It doesn’t sort everything in one flashy move,but in...

[Algorithm] Quick Sort — Divide and Conquer!

⚡ Quick Sort: Divide and Conquer! “Divide each difficulty into as many parts as is feasible and necessary to resolve it.”— René Descartes In life, as in sorting algorithms, you don’t have to solve a complex problem all at once.Break it down into smaller pieces, and it becomes much easier to handle.Quick Sort is an algorithm built exactly on that philosophy. Imagine organizing your bookshelf.You pick one book in the middle as a reference point —thinner books go to the left, thicker book...

[Algorithm] Selection Sort — The Power of Choosing

🔍 What’s the Big Deal About Selection Sort? “Destiny is not a matter of chance, it is a matter of choice.” – William Jennings Bryan In life, we constantly face moments of choice.Who will you walk with? What will you do first?Ultimately, these choices shape the course of your life. In programming, there’s an algorithm that embodies this same idea —Selection Sort. Selection Sort repeatedly finds the smallest value in the list,and moves it to the front, step by step, until the entire lis...