Dataset

yelp_restaurant_reviews

News & Sentiment Download 1 source

This dataset is a filtered version of https://huggingface.co/datasets/vincha77/filtered_yelp_restaurant_reviews

Open on Hugging Face

Data preview

Subset

Loading preview...

Dataset card

yelp_restaurant_reviews is a small filtered sample of Yelp restaurant reviews published on Hugging Face by bespokelabs, distributed in Parquet format with sentiment labels for natural language processing work. The dataset contains 1,000 rows in a single train split and is tagged for use with the datasets, pandas, mlcroissant and polars libraries.

What this dataset contains

The dataset is a filtered subset of Yelp restaurant reviews, consisting of 1,000 text rows in a train split. Each row carries a sentiment label drawn from a fixed three-class taxonomy of negative, neutral, and positive. Review length is also recorded as an integer field, allowing length-based analysis alongside the raw review text.

Structure and fields

The Parquet file exposes three columns. The `text` column holds the raw review string, the `label` column holds a categorical sentiment class from the set {negative, neutral, positive}, and the `review_length` column stores an int64 character count for each review. The single train split contains 1,000 rows and the file is loadable through datasets, pandas, mlcroissant or polars.

How analysts use it

Analysts typically use the dataset for sentiment-classification benchmarking and as a lightweight text corpus for prototyping NLP pipelines. The three-class label scheme makes it suitable for evaluating consumer-attitude scoring models that could in principle be applied to brand or sector monitoring, though the 1,000-row size limits statistical significance for any production signal. It can also serve as a teaching example for label distribution and review-length analysis.

Documentation & references 1