Amazon Product Dataset

Column Descriptions

• ASIN: A unique identifier assigned to each product on Amazon (Amazon Standard Identification Number).
• Product Title: The name or title of the product.
• Current Price: The most recent price of the product.
• Original Price: The price of the product before any discounts were applied.
• Currency: The currency used for the product’s pricing.
• Star Rating: The average customer rating of the product based on a 5-star system.
• Number of Ratings: The total number of ratings the product has received from customers.
• Product URL: The web address linking directly to the product’s page on Amazon.
• Product Image URL: A link to an image or photo of the product.
• Number of Offers: The number of distinct offers or sellers available for the product.
• Minimum Offer Price: The lowest price among the available offers for the product.
• Best Seller Status: Whether the product is listed as a best seller on Amazon.
• Amazon Choice Status: Whether the product is designated as Amazon’s Choice.
• Prime Eligibility: Whether the product is eligible for Amazon Prime benefits.
• Climate Pledge Friendly: Indicates if the product carries the Climate Pledge Friendly label.
• Sales Volume: The estimated number of units sold for the product.
• Delivery Options: Information on delivery services and time frames for the product.
• Product Variations: A flag indicating whether the product has variations like different sizes or colors.
• Availability: The stock status or availability of the product.
• Unit Price: The cost per unit of measurement for the product.
• Unit Count: The quantity of units included in the listed price.

Potential Analyses and Data Science Applications

Descriptive Analytics:

• Perform summary statistics on numerical columns like prices, ratings, and sales volumes (e.g., mean, median, range).
• Frequency analysis for categorical columns such as best seller status, Amazon Choice, and Prime eligibility.

Price Trend Analysis:

• Compare current prices to original prices to assess discounts.
• Identify pricing trends across product categories, brands, or sellers.

Rating and Review Analysis:

• Examine the distribution of product ratings and correlate them with the number of ratings to identify trends.
• Study how product ratings relate to features like price and sales volume.

Sales Volume Analysis:

• Identify the top-selling products by sales volume.
• Analyze how sales volumes relate to pricing, ratings, and other attributes.

Product Segmentation:

• Group products based on attributes like best seller status, Amazon’s Choice, Prime eligibility, or Climate Pledge Friendly designation.
• Use clustering algorithms to uncover patterns or segments within the product offerings.

Predictive Modeling:

• Price Prediction: Develop regression models (e.g., linear regression, decision trees) to predict prices based on variables such as star ratings, number of offers, and whether a product is a best seller.
• Sales Volume Prediction: Use regression or time series models to predict future sales volumes based on historical data.
• Rating Prediction: Predict product ratings using features like price, number of ratings, and seller status.

Recommendation Systems:

• Build recommendation models (e.g., collaborative filtering, content-based filtering) to suggest products based on user preferences or purchase history.

Classification Tasks:

• Classify products into categories like best seller, Amazon’s Choice, or Prime eligible using classification algorithms such as logistic regression, random forests, or support vector machines (SVM).

Sentiment and Review Analysis:

• Perform sentiment analysis on customer reviews (if available) and correlate the results with product ratings and sales volumes.

Market Basket Analysis:

• If transaction data is available, use association rule mining techniques to identify frequently co-purchased products.

Visualization Techniques

Data Visualization:

• Histograms & Bar Charts: Visualize distributions of numeric columns like prices, ratings, and sales volumes.
• Box Plots: Compare product prices and ratings across different categories.
• Scatter Plots: Analyze relationships between continuous variables (e.g., price vs. sales volume).
• Heatmaps: Display correlations between different attributes.

Data Preprocessing

Data Cleaning:

• Handle missing data by either imputing values or removing incomplete records.
• Convert categorical variables into numeric values using techniques like one-hot encoding.

Scaling & Normalization:

• Scale numerical features for algorithms that are sensitive to the distribution of data (e.g., SVM or k-means clustering).

Advanced Techniques

Feature Engineering:

• Create new features, such as calculating the discount percentage from the original and current prices.

Dimensionality Reduction:

• Use methods like Principal Component Analysis (PCA) to reduce dataset dimensionality if necessary.