Islamophobic Hate Speech on Social Media with Weak and Strong Labels

Bertie Vidgen at The Alan Turing Institute released this dataset in 2019 to support research on detecting Islamophobic hate speech. The data includes social media text annotated for both weak and strong hate speech categories. The release also includes the associated code and annotation guidelines from the original publication.

Use Cases

• Train classifiers to detect Islamophobic hate speech based on annotated social media text.
• Study the linguistic features distinguishing weak and strong hate speech categories.
• Benchmark model performance on a specific, annotated hate speech detection task.
• Apply or adapt the provided annotation guidelines for labeling similar social media content.

Strengths

• Includes annotation guidelines, which provide transparency into the labeling process.
• Focuses on a specific, nuanced classification task of weak versus strong hate speech.
• Associated code from the original 2019 publication is included.

Limitations

• Row count and dataset size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Bertie Vidgen, The Alan Turing Institute

Collection Method

Social media text annotated for a research publication.