Data Acquisition and Exploration on Amazon Sales Dataset

A comprehensive analysis focusing on data acquisition, exploration, and preliminary insights derived from the Amazon Sales Dataset.

Table of Contents

• Dataset Overview
• Task
• Dataset Features
• Visualization Highlights
• Dataset Quality Assessment
• Potential Machine Learning Implementations

Dataset Overview

• Dataset Source: Amazon Sales Dataset on Kaggle

• Description:

  • The dataset covers 1465 entries with 16 distinct attributes, detailing authentic Amazon Product Ratings and Reviews.
  • The data has been extracted from Amazon's official website, containing 1351 unique values and 1194 unique user IDs.
  • It serves as a strong tool for business analytics, especially for retail and e-commerce domains.
  • The primary focus of this project is on data acquisition, exploration, and foundational analysis.

Task

Amazon, a leading multinational e-commerce company, offers a various of products and services globally. This dataset captures product information and real customer reviews, enabling researchers and analysts to reveal market trends, consumer behavior, and sales insights.

Dataset Features

• product_id: Unique identifier for each product.
• product_name: Descriptive name of the product.
• category: Classification category of the product.
• discounted_price: Offered price after discounts.
• actual_price: Original price of the product without discounts.
• discount_percentage: Discount rate applied to the product.
• rating: Average rating given to the product.
• rating_count: Number of individual ratings the product received.
• about_product: Brief description of the product.
• user_id: Unique identifier for each user.
• user_name: Name of the user who provided the review.
• review_id: Unique identifier for each review.
• review_title: Short summary of the review.
• review_content: Detailed review content.
• img_link: Link directing to the product's image.
• product_link: Direct link to the official product page.

Visualization Highlights

1. Distribution of Top 10 Product Categories Sales:

   • A pie chart representation revealing the significant sales of computer and electronic products, capturing 36.2% and 11.9% of sales, respectively.

   • The dataset identifies 211 distinct categories, offering insights into the highest and lowest contributors in sales.

   • Amazon appears as a prime marketplace for customers seeking electronic accessories.

     

2. Average Product Rating of 50 Categories:

   • The bar chart suggests the Computer & Accessories category as the frontrunner with an average rating of approximately 4.6/5.0.

   • The graph subtly indicates the popularity and demand for computer and electronic accessories.

     

3. Top 10 Best-Selling Products With The Highest Rating Counts:

   • AmazonBasics Flexible Premium HDMI cable is recognized as the top-rated product with over 800,000 ratings globally.
   • A recurring trend is observed with computer and electronic accessories reigning in popularity and demand.

Dataset Quality Assessment

• Data Validity:

  • The dataset is authenticated and ratified by Kaggle.
  • Each record references the official product page and corresponding image, underscoring the dataset's authenticity.

• Data Accessibility:

  • Accessible to all under the CC BY-NC-SA 4.0 license.
  • The dataset is well-structured, facilitating smooth data manipulations and visualizations.

• Data Completeness:

  • While the dataset is comprehensive, minor missing values in the ratings column were addressed during the exploration phase.

• Data Consistency:

  • Some discrepancies in data formatting and structure were noted, yet the dataset predominantly maintains consistency.

• Data Size:

  • The dataset, with its 1465 entries, is relatively concise, but it provides substantial data for preliminary analysis.

Potential Machine Learning Implementations

1. Price Optimization:

   • A predictive model to refine pricing strategies based on market dynamics and consumer feedback.

2. Product Recommendation:

   • A machine learning-driven recommendation system suggesting products aligned with a user's purchase history and reviews.

3. Sentiment Analysis:

   • A model to sift through product reviews and categorize them as positive, negative, or neutral, offering insights into customer satisfaction and product areas requiring enhancement.

Note: The dataset, though insightful, presents certain limitations like sample size, geographical bias, and data recency. These factors should be considered when extrapolating results or venturing into extensive machine learning applications.