Pepere45/spanish-wikipedia-corpus

Spanish Wikipedia Corpus

Historical Spanish-language content collected via the Wayback Machine (Internet Archive).

Dataset Summary

• Language: Spanish (es)
• Sources: es.wikipedia.org — history, science, culture, geography, politics
• Period: 2005–2023
• Topics: historia, ciencia, cultura, geografía, política...
• License: Public archive snapshots. Verify rights before commercial use.

Features

Each record includes automatic labeling:

• topics: detected topic tags
• region: inferred geographic region (espana_general, cataluna, andalucia, madrid...)
• era: linguistic era (web_1_0, pre_social, social_media, movil_first, ia_era)
• sentiment: positivo / neutro / negativo
• sentiment_score: sentiment intensity -1.0 to +1.0
• quality_score: document quality 0–100
• readability_score: Flesch readability adapted for Spanish
• lexical_density: vocabulary richness 0–1
• word_count: total word count
• entities: detected flags (has_prices, has_dates, has_quotes...)
• content_hash: MD5 deduplication fingerprint
• captured_at: original capture date
• source_domain: origin website
• wayback_url: Internet Archive snapshot URL

Use Cases

• High-quality Spanish language modeling
• Knowledge base training for Spanish LLMs
• Encyclopedic Spanish NLP research
• Cross-domain Spanish text analysis

Data Collection

Content was collected from the Internet Archive's Wayback Machine using an automated pipeline with:

• Respectful crawling delays (2–6 seconds between requests)
• Automatic quality filtering (minimum quality score 70/100)
• MD5-based deduplication
• Spanish language detection

Contact

For bulk licensing, custom datasets or research collaborations: contact via Hugging Face profile.

Citation

If you use this dataset in your research, please cite:

@dataset{pepere45_spanish_wikipedia_corpus,
  author = {Pepere45},
  title = {Spanish Wikipedia Corpus},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/Pepere45/spanish-wikipedia-corpus}
}