undertheseanlp/UVB-v0.1

UVB - Underthesea Vietnamese Books Dataset

A collection of 447 Vietnamese books with full text content and Goodreads metadata for NLP research.

Dataset Summary

UVB (Underthesea Vietnamese Books) is a dataset containing 447 Vietnamese books with full text content, mapped to Goodreads for metadata enrichment including genres, ratings, and publication years. The dataset is designed for Vietnamese language model training, text generation, and other NLP tasks.

Supported Tasks

• Language Modeling: Pre-training or fine-tuning Vietnamese language models
• Text Generation: Training models to generate Vietnamese text
• Text Classification: Using genres as labels for classification tasks
• Information Extraction: Extracting structured information from book texts

Languages

Vietnamese (vi)

Dataset Structure

Data Instances

{
  "id": "vn_000001",
  "title": "The DIGITAL ECONOMY 20th Anniversary Edition",
  "author": "Don Tapscott",
  "content": "The DIGITAL ECONOMY 20th Anniversary Edition Rethinking Promise and Peril...",
  "genres": ["Business", "Paradigm Shift"],
  "first_publish": "2014",
  "goodreads_id": "20722002-the-digital-economy",
  "goodreads_url": "https://www.goodreads.com/book/show/20722002-the-digital-economy",
  "goodreads_rating": 4.1,
  "goodreads_num_ratings": 29
}

Data Fields

Field	Type	Description
id	string	Unique identifier (e.g., vn_000001)
title	string	Book title
author	string	Author name
content	string	Full text content of the book
genres	list[string]	Book genres from Goodreads
first_publish	string	First publication year
goodreads_id	string	Goodreads book ID
goodreads_url	string	Goodreads URL
goodreads_rating	float	Goodreads rating (1-5)
goodreads_num_ratings	int	Number of ratings on Goodreads

Data Splits

Split	Examples
train	447

Dataset Statistics

• Total books: 447
• Books with genres: 230 (51.5%)
• Books with publication year: 421 (94.2%)
• Total size: ~209 MB

Top Genres:

Genre	Count
Non Fiction	76
Fiction	62
Romance	37
Classics	30
Novels	27
Philosophy	25
Self Help	25

Publication Year Distribution:

Period	Count
Before 1900	6
1900-1950	12
1951-1980	38
1981-2000	82
2001-2010	134
2011+	149

Dataset Creation

Source Data

• Vietnamese books: tmnam20/Vietnamese-Book-Corpus
• Goodreads metadata: BrightData/Goodreads-Books

Data Collection

1. Vietnamese books were loaded from the source corpus
2. Books were matched to Goodreads entries using fuzzy title/author matching
3. Genres and publication years were scraped from Goodreads pages
4. Data was merged into a single dataset with full text content

Processing Scripts

Processing scripts are included in the scripts/ directory:

• map_goodreads.py - Map Vietnamese books to Goodreads entries
• add_genres.py - Fetch genres from Goodreads
• add_publish_date.py - Fetch first publication year
• requirements.txt - Python dependencies

Usage

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("undertheseanlp/UVB-v0.1")

# Access the data
for item in dataset["train"]:
    print(f"Title: {item['title']}")
    print(f"Author: {item['author']}")
    print(f"Content: {item['content'][:200]}...")
    print(f"Genres: {item['genres']}")
    print(f"First publish: {item['first_publish']}")
    break

# Filter by genre
fiction_books = [item for item in dataset["train"] if "Fiction" in (item.get("genres") or [])]
print(f"Found {len(fiction_books)} fiction books")

Considerations for Using the Data

Social Impact

This dataset can help advance Vietnamese NLP research and applications, contributing to language technology for Vietnamese speakers.

Limitations

• Book content may contain OCR errors from digitization
• Genre and metadata coverage is not 100%
• Some author names may include translator information

Additional Information

Licensing

Apache 2.0

Citation

@misc{uvb_dataset,
  title={UVB: Underthesea Vietnamese Books Dataset},
  author={Underthesea NLP},
  year={2026},
  publisher={HuggingFace},
  url={https://huggingface.co/datasets/undertheseanlp/UVB-v0.1}
}

Contributions

Thanks to @rain1024 for creating this dataset.