🛒 E-Commerce Customer Data Cleaning & EDA Project

📌 Project Overview

This project performs end-to-end data cleaning, exploratory data analysis (EDA), visualization, and data transformation on a messy e-commerce sales dataset using Python. The goal is to take raw, inconsistent data and turn it into a clean, analysis-ready dataset while uncovering key patterns in customer purchasing behavior.

🎯 Objective

To clean and analyze an E-commerce Customer dataset using Python and perform Exploratory Data Analysis (EDA) to extract meaningful business insights.

📂 Dataset

• Input file: messy_ecommerce_sales_data.csv
• Output file: Cleaned_Ecommerce_Data.csv (generated after cleaning)
• Columns: ID, Customer_Name, Order_ID, Order_Date, Product, Category, Quantity, Price, Payment_Method, Status, Total

Note: If the raw dataset isn't your own to redistribute, consider linking to its original source instead of uploading it directly.

🧹 Data Cleaning Steps

Each column was cleaned individually following an identify → solve → verify workflow:

Column	Issues Found	Fix Applied
ID	Duplicate values	Removed duplicate rows
Customer_Name	Duplicate values	Removed duplicate rows
Order_ID	None	No cleaning needed
Order_Date	Invalid formats, missing values	Converted to datetime, filled missing with mode, standardized format (dd-mm-yyyy)
Product	Inconsistent casing	Converted to title case
Category	Inconsistent casing, spelling variants, missing values	Standardized text, merged spelling variants (e.g. "Electronic" → "Electronics"), filled missing with mode
Quantity	Invalid/negative values	Converted to numeric, replaced negatives with NaN, filled missing with mode
Price	Currency symbols, invalid/negative values, outlier	Removed $ symbol, converted to numeric, replaced negatives with NaN, filled missing with median, removed outlier row (price = 10000)
Payment_Method	None	No cleaning needed
Status	None	No cleaning needed
Total	Negative values, missing values	Replaced negatives with NaN, filled missing with median

📊 Exploratory Data Analysis & Visualization

The following visualizations were created to explore relationships in the cleaned data:

• Scatter Plot — Quantity vs Total
• Bar Chart — Order count by Payment Method
• Horizontal Bar Chart — Total sales by Payment Method
• Pie Chart — Order distribution by Category
• Seaborn Bar Plot — Price by Category and Payment Method
• Box Plot — Quantity distribution by Category
• Correlation Heatmap — Price, Quantity, and Total

Key Insights

• Price and Total show a moderate positive correlation (0.64)
• Quantity and Total show a moderate positive correlation (0.58)
• Price and Quantity have a very weak correlation (0.07)
• Total is mainly driven by both Price and Quantity together

🔄 Data Transformation

• Label Encoding using sklearn.preprocessing.LabelEncoder
• One-Hot Encoding using pandas.get_dummies() on Payment_Method and Status columns (with and without drop_first)

🛠️ Tools & Libraries

• Python 3
• pandas
• numpy
• matplotlib
• seaborn
• scikit-learn (LabelEncoder)

🚀 How to Run

1. Clone this repository

   git clone <your-repo-url>
   cd <repo-folder>

2. Install dependencies

   pip install pandas numpy matplotlib seaborn scikit-learn

3. Open the notebook

   jupyter notebook Home_Based_Project.ipynb

   or open it directly in Google Colab.

📁 Project Structure

ecommerce-customer-eda/
│
├── Home_Based_Project.ipynb        # Main notebook (cleaning, EDA, visualization, transformation)
├── data/
│   ├── messy_ecommerce_sales_data.csv    # Raw input dataset (optional to include)
│   └── Cleaned_Ecommerce_Data.csv        # Cleaned output dataset
├── images/                          # (optional) exported chart screenshots for the README
├── README.md
└── requirements.txt                  # Python dependencies

👤 Author

Maira Rehmat ID: AI-817677

📄 License

This project is open source and available for educational use.