reviewsAmazon

34,686,770 Amazon reviews from 6,643,669 users

Dataset Overview

This dataset encompasses a vast collection of 34,686,770 Amazon reviews from 6,643,669 users across 2,441,053 products, offering a deep dive into consumer sentiment.

Data Origin

Originating from the Stanford Network Analysis Project (SNAP), this data spans 18 years up to March 2013, providing an extensive historical view of consumer opinions.

Sentiment Classification

Reviews are classified into negative and positive sentiments based on their scores, with 1-2 considered negative and 4-5 positive, excluding neutral reviews.

Dataset Structure

The polarity dataset is meticulously structured, containing 1,800,000 training samples and 200,000 testing samples for each sentiment category.

File Format

Available in CSV format, each file contains polarity, title, and text columns, representing the sentiment class, review heading, and review body respectively.

Usage Guidelines

Researchers and developers can leverage this dataset to analyze trends, perform sentiment analysis, and develop machine learning models to understand consumer behavior.

TOKEN SHOWCASE

List of tokens people are building with Solana

🙏 Please add your token

BTC