Spanish Offensive Comments From Young Influencers

30,416 Spanish social media posts manually labeled into offensive categories. It focuses on comments from young influencers on Twitter, Instagram, and YouTube and was used in the MeOffendES competition at IberLEF 2021.

Use Cases

• Train a text classifier to predict offensive categories for Spanish social media posts.
• Analyze the distribution of offensive language across different pre-defined categories.
• Benchmark models for the specific task of Spanish offensive language detection.

Strengths

• Contains 30,416 manually labeled posts.
• Labels are based on pre-defined offensive categories.
• Data is sourced from major social platforms: Twitter, Instagram, and YouTube.

Limitations

• The specific column structure and features are unknown.
• The sample represents a subset selected for a competition, which may introduce bias.
• The geographic origin of the comments within the Spanish-speaking world is unspecified.

Provenance

Source

Social media platforms (Twitter, Instagram, YouTube).

Collection Method

Manually collected and labeled corpus.

Freshness

Last updated on 2024-03-22.