|
|
--- |
|
|
license: mit |
|
|
task_categories: |
|
|
- text-classification |
|
|
- text-generation |
|
|
- summarization |
|
|
- feature-extraction |
|
|
language: |
|
|
- en |
|
|
size_categories: |
|
|
- 1K<n<10K |
|
|
tags: |
|
|
- news |
|
|
- articles |
|
|
- bbc |
|
|
- nlp |
|
|
- text |
|
|
pretty_name: BBC News Articles Dataset |
|
|
--- |
|
|
|
|
|
# BBC News Articles Dataset |
|
|
|
|
|
## Dataset Description |
|
|
|
|
|
A collection of **2,225 news articles** from BBC, suitable for text classification, summarization, and NLP tasks. |
|
|
|
|
|
### Dataset Summary |
|
|
|
|
|
| Metric | Value | |
|
|
|--------|-------| |
|
|
| **Total Articles** | 2,225 | |
|
|
| **Unique Articles** | 2,092 | |
|
|
| **Columns** | `filename`, `article_text` | |
|
|
| **Language** | English | |
|
|
| **Source** | BBC News | |
|
|
|
|
|
## Dataset Structure |
|
|
|
|
|
### Data Fields |
|
|
|
|
|
| Field | Type | Description | |
|
|
|-------|------|-------------| |
|
|
| `filename` | string | Unique identifier/filename for each article | |
|
|
| `article_text` | string | Full text content of the news article | |
|
|
|
|
|
### Text Statistics |
|
|
|
|
|
| Metric | Min | Max | Mean | Median | Std | |
|
|
|--------|-----|-----|------|--------|-----| |
|
|
| **Characters** | 470 | 25,453 | 2,232 | 1,935 | 1,364 | |
|
|
| **Words** | 84 | 4,428 | 379 | 326 | 238 | |
|
|
| **Sentences** | 4 | 248 | 19 | 16 | 13 | |
|
|
|
|
|
|
|
|
|
|
|
### Vocabulary Statistics |
|
|
|
|
|
| Metric | Value | |
|
|
|--------|-------| |
|
|
| **Total Words (corpus)** | 815,279 | |
|
|
| **Unique Words (vocabulary)** | 27,205 | |
|
|
| **Vocabulary (excl. stopwords)** | 27,070 | |
|
|
| **Lexical Diversity** | 0.0334 | |
|
|
| **Avg Words per Article** | 366.4 | |
|
|
|
|
|
|
|
|
### Top 10 Most Frequent Words |
|
|
|
|
|
| Word | Frequency | |
|
|
|------|-----------| |
|
|
| said | 7,253 | |
|
|
| mr | 2,994 | |
|
|
| would | 2,628 | |
|
|
| also | 2,156 | |
|
|
| people | 2,041 | |
|
|
| new | 1,898 | |
|
|
| us | 1,818 | |
|
|
| year | 1,813 | |
|
|
| one | 1,752 | |
|
|
| could | 1,534 | |
|
|
|
|
|
|
|
|
 |
|
|
|
|
|
## Usage |
|
|
|
|
|
### Loading with Hugging Face Datasets |
|
|
|
|
|
```python |
|
|
from datasets import load_dataset |
|
|
|
|
|
dataset = load_dataset("Omarrran/BBC_Eng_News_Articles_dataset") |
|
|
|
|
|
# Access training data |
|
|
train_data = dataset['train'] |
|
|
|
|
|
# View first article |
|
|
print(train_data[0]['article_text'][:500]) |
|
|
``` |
|
|
|
|
|
### Loading with Pandas |
|
|
|
|
|
```python |
|
|
import pandas as pd |
|
|
from datasets import load_dataset |
|
|
|
|
|
dataset = load_dataset("Omarrran/BBC_Eng_News_Articles_dataset") |
|
|
df = dataset['train'].to_pandas() |
|
|
|
|
|
# Basic exploration |
|
|
print(f"Total articles: {len(df)}") |
|
|
print(df.head()) |
|
|
``` |
|
|
|
|
|
### Text Classification Example |
|
|
|
|
|
```python |
|
|
from sklearn.model_selection import train_test_split |
|
|
from sklearn.feature_extraction.text import TfidfVectorizer |
|
|
from sklearn.naive_bayes import MultinomialNB |
|
|
|
|
|
# Load data |
|
|
df = dataset['train'].to_pandas() |
|
|
|
|
|
# If categories available from filename |
|
|
df['category'] = df['filename'].apply(lambda x: x.split('/')[0]) |
|
|
|
|
|
# Split |
|
|
X_train, X_test, y_train, y_test = train_test_split( |
|
|
df['article_text'], df['category'], test_size=0.2, random_state=42 |
|
|
) |
|
|
|
|
|
# Vectorize and train |
|
|
vectorizer = TfidfVectorizer(max_features=5000, stop_words='english') |
|
|
X_train_vec = vectorizer.fit_transform(X_train) |
|
|
X_test_vec = vectorizer.transform(X_test) |
|
|
|
|
|
clf = MultinomialNB() |
|
|
clf.fit(X_train_vec, y_train) |
|
|
print(f"Accuracy: {clf.score(X_test_vec, y_test):.2%}") |
|
|
``` |
|
|
|
|
|
### Summarization Example |
|
|
|
|
|
```python |
|
|
from transformers import pipeline |
|
|
|
|
|
summarizer = pipeline("summarization", model="facebook/bart-large-cnn") |
|
|
|
|
|
# Summarize first article |
|
|
article = dataset['train'][0]['article_text'] |
|
|
summary = summarizer(article[:1024], max_length=130, min_length=30) |
|
|
print(summary[0]['summary_text']) |
|
|
``` |
|
|
|
|
|
## Suitable Tasks |
|
|
|
|
|
This dataset is ideal for: |
|
|
|
|
|
- **Text Classification**: Categorize articles by topic |
|
|
- **Summarization**: Generate article summaries |
|
|
- **Named Entity Recognition**: Extract entities from news |
|
|
- **Keyword Extraction**: Identify key topics |
|
|
- **Topic Modeling**: Discover latent themes |
|
|
- **Sentiment Analysis**: Analyze article tone |
|
|
- **Text Generation**: Fine-tune language models |
|
|
- **Information Retrieval**: Build search systems |
|
|
|
|
|
## Data Quality |
|
|
|
|
|
| Check | Status | |
|
|
|-------|--------| |
|
|
| Empty/null articles | 0 found | |
|
|
| Encoding issues | Clean (UTF-8) | |
|
|
|
|
|
## Limitations |
|
|
|
|
|
- Dataset is limited to BBC News articles |
|
|
- May contain temporal bias based on collection period |
|
|
- English language only |
|
|
- News domain specific vocabulary |
|
|
|
|
|
## Citation |
|
|
|
|
|
```bibtex |
|
|
@dataset{bbc_news_articles, |
|
|
title = {BBC_Eng_News_Articles_dataset_hnm}, |
|
|
Author ={Haq Nawaz Malik} |
|
|
year = {2025}, |
|
|
publisher = {Hugging Face}, |
|
|
howpublished = {\url{https://huggingface.co/datasets/Omarrran/BBC_Eng_News_Articles_dataset/}} |
|
|
} |
|
|
``` |
|
|
|
|
|
## License |
|
|
|
|
|
This dataset is provided for research and educational purposes under the MIT License. |
|
|
|
|
|
--- |
|
|
|
|
|
|
|
|
|
