Skip to content

Data cleaning and exploratory analysis of used car listings from eBay Kleinanzeigen to identify brand trends, pricing patterns, and mileage behavior.

License

Notifications You must be signed in to change notification settings

linetrono/Used-Car-Pricing-Analysis-Python_Project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Cleaning and Exploring eBay Car Sales Data

This project involves cleaning and exploring used car sales data that originated from eBay-Kleinanzeigen (a German classifieds site).
The main goal was to perform data cleaning, and identify pricing patterns, brand trends, and mileage behaviors that can help potential buyers or sellers make data-driven decisions in the automotive market.


Objectives

  • Clean and prepare raw car listing data for analysis.
  • Explore trends and relationships between variables (e.g., price vs. mileage, brand vs. price).
  • Identify outliers and unrealistic listings.
  • Derive insights into the most cost-efficient brands and models.
  • Visualize data to highlight key pricing trends.

Dataset Information

Source: eBay Kleinanzeigen Used Car Listings Dataset via Dataquest (German used car marketplace — eBay Kleinanzeigen)

Description: The dataset contains information on used car listings, including:

  • Price
  • Brand
  • Vehicle type
  • Registration year
  • Power (PS)
  • Mileage (km)
  • Date listed

Key Steps Performed

  • Removed inaccurate and unrealistic values
  • Converted data types (e.g., numeric columns stored as text)
  • Normalized column names
  • Filtered data to realistic price & mileage ranges
  • Performed exploratory data analysis (EDA)

Key Insights

  • Volkswagen is the most frequently listed brand even though its average price sits around the mid-range; strongest market share.
  • Premium German brands (BMW, Mercedes, Audi) have significantly higher average prices, but not significantly lower mileage; pricing is driven more by brand equity than remaining vehicle life.
  • Budget brands (Ford, Opel) have lower average prices and similar mileage ranges → good value for buyers with tighter budgets.
  • Mileage variation across top brands is relatively narrow (≈ within ±10%); mileage is not the primary driver of price in this dataset.
  • Volkswagen balances value and demand
  • Audi is the priciest on average (in this dataset), positioning it at the top of the premium segment.

Recommendations

  • Buyers looking for the best balance between price and reliability perception should consider Volkswagen.
  • Premium buyers prioritizing luxury experience and brand prestige can target BMW / Mercedes / Audi, but should not expect better mileage for the higher price.
  • Price-sensitive buyers can choose Ford or Opel for good affordability without a major sacrifice in typical mileage.
  • Used-car buyers should not assume “high price = low mileage”; instead they should compare specific car history (service records, accident history, previous owners) not just brand label.

Tools & Technologies Used

  • Python
  • Pandas for data cleaning and manipulation
  • NumPy for numerical operations
  • Jupyter Notebook for exploration and presentation

Files

Analyis Notebook; Jupyter Notebook: Data Cleaning & EDA

CSV File: Ebay Dataset


How to Use

1. Clone the Repository

git clone https://github.com/linetrono/Used-Car-Pricing-Analysis-Python_Project.git
cd Used-Car-Pricing-Analysis-Python_Project

2. Install Dependencies

Ensure you have Python 3 installed, then install required libraries: Numpy & Pandas

pip install pandas numpy 

3. Run the Notebook

jupyter notebook notebooks/ebay_car_sales_cleaning_eda.ipynb

Contact

For questions, collaboration, or feedback, feel free to reach out:

Email:

LinkedIn:

GitHub:

About

Data cleaning and exploratory analysis of used car listings from eBay Kleinanzeigen to identify brand trends, pricing patterns, and mileage behavior.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published