NapoleonBonaprt/amazon-egypt-ecommerce-analysis

# Amazon Egypt E-Commerce Analysis ## 📌 Project Overview This project analyzes an Amazon Egypt e-commerce product dataset to understand product characteristics, customer engagement, pricing, discounts, and factors associated with Best Seller status. The project combines **Data Preparation, Exploratory Data Analysis (EDA), Machine Learning, Power BI Visualization, and AI-Assisted Insights**. ## 🎯 Objectives - Clean and prepare real-world e-commerce data. - Analyze pricing, discounts, ratings, reviews, purchases, and product categories. - Identify patterns related to Best Seller products. - Build and compare classification models to predict Best Seller status. - Create an interactive Power BI dashboard. - Translate analytical findings into actionable business recommendations. ## 📊 Dataset - **Source:** Amazon Egypt product listings collected through web scraping - **Business Domain:** E-commerce / Retail / Product Analytics - **Original Size:** Approximately 25,000 rows and 14 columns - **After Preparation:** Approximately 21,250 product records and 15 variables - **Target Variable:** `is_best_seller_flag` - `1` = Best Seller - `0` = Not Best Seller ### Main Features - Product Category - Original Price - Discounted Price - Discount Percentage - Product Rating - Total Reviews - Purchases Last Month - Sponsored Status - Coupon - Delivery Date - Best Seller Status ## 🧹 Data Preparation Data preparation was performed using **Python, Pandas, and NumPy**. The workflow included: - Inspecting data types and missing values - Removing duplicate records - Standardizing inconsistent values and formats - Converting numerical fields to appropriate data types - Validating price relationships - Detecting and treating outliers - Cleaning the Best Seller target variable - Encoding categorical variables - Applying feature scaling where appropriate for Machine Learning - Keeping an unscaled dataset for Power BI After cleaning, invalid price relationships were reduced to **0 …