SpecificityTwitter: Censored Tweets with Language Specificity Annotations

Twitter posts annotated for language specificity, developed by researcher cs329yangzhong for a 2019 AAAI publication. It provides a collection of social media text where each entry is evaluated based on how specific or general the language is, though the exact record count is not specified in the repository metadata.

Use Cases

• Training NLP models to predict specificity levels in short-form social media text
• Analyzing linguistic patterns to distinguish between vague and detailed user communication
• Benchmarking specificity-aware language models against the AAAI 2019 paper results

Strengths

• Annotated based on a peer-reviewed AAAI 2019 research framework
• Targets language specificity, a niche linguistic feature distinct from sentiment or topic
• Derived from real-world social media mining

Limitations

• Tweets are censored, which may limit the ability to perform certain contextual or entity-linkage NLP tasks
• Lack of metadata regarding the total row count or class distribution

Provenance

Source

AAAI 2019 paper: Predicting and Analyzing Language Specificity in Social Media Posts

Collection Method

annotated

Freshness

The dataset was last updated in October 2021 and is based on research published in 2019.