📊 Amazon Product Reviews - Exploratory Data Analysis (EDA)

📌 Overview

This project performs Exploratory Data Analysis (EDA) on an Amazon product dataset.
The dataset contains product details, prices, discounts, ratings, reviews, and user information.

The goal of this analysis is to:

• Understand the structure and quality of the dataset.
• Identify trends in pricing, discounting, and ratings.
• Explore customer review patterns.
• Detect potential issues like missing values, duplicates, or imbalances.

---

🗂️ Dataset Description

The dataset includes the following key columns:

Column	Description
product_id	Unique identifier for each product
product_name	Name/description of the product
category	Product category (e.g., Electronics, Accessories)
discounted_price	Selling price after discount
actual_price	Original price before discount
discount_percentage	Percentage discount offered
rating	Customer rating (out of 5)
rating_count	Number of ratings
about_product	Short description/features
user_id	Unique ID of reviewer
user_name	Name of reviewer
review_id	Unique ID of review
review_title	Title of review
review_content	Full review text
img_link	Product image link
product_link	Product page link

---

🔍 Steps in EDA

1. Data Inspection

• Used .info() to check data types, null values, and dataset size.
• Found that most columns are complete, with very few missing values.

2. Descriptive Statistics

• .describe() applied to both numeric and categorical columns.
• Found mean ≈ median in prices → data is fairly symmetric.
• Ratings cluster around 4.1, showing positive bias.

3. Correlation Analysis

• Computed correlation matrix for numeric features.
• Observed strong negative correlation between discount_percentage and discounted_price.
• Weak/no correlation between rating and price → ratings are not price-driven.

4. Visualizations

• Bar Chart: Average rating per category.
• Boxplot: Discount % distribution across categories.
• Scatterplot: Discounted price vs rating.
• Word Cloud: Most frequent terms in reviews.
• Heatmap: Correlations between numeric features.

5. Data Quality Checks

• Found duplicate product IDs (same product reviewed multiple times).
• Prices and discounts stored as strings (₹, %) → cleaned and converted to numeric.

---

📈 Insights

• Many products receive 4★ or higher → customer reviews skew positive.
• Discounts are widely offered (~50% most frequent).
• Certain categories dominate the dataset (e.g., Electronics & Accessories).
• Some reviews and users appear multiple times → dataset contains duplicate/overlapping entries.

---

🛠️ Tools & Libraries

• Python 3
• Pandas → data cleaning & manipulation
• NumPy → numerical operations
• Matplotlib / Seaborn → data visualization
• WordCloud → review text analysis

---

📌 How to Run

1. Clone the repository:

   git clone https://github.com/HarshitWaldia/Exploratory-Data-Analysis.git
   cd Exploratory-Data-Analysis
   

2. Install required libraries:

pip install -r requirements.txt

3.Open the Jupyter Notebook:

jupyter notebook Amazon_EDA.ipynb

4. Run the cells step by step to reproduce the analysis.

🚀 Future Work

• Build a recommendation system using ratings & categories.

• Perform sentiment analysis on review text.

• Use ML models to predict product ratings based on price & discount.

👨‍💻 Author

Harshit Waldia