Dz-Emotion: 6,000 Algerian Arabic Social Media Comments for Emotion Detection

6,000 manually annotated social media comments collected from YouTube, Facebook, and Instagram. The dataset, Dz-Emotion, is the first large-scale resource for emotion detection in the Algerian Arabic dialect (Darija), labeled according to Ekman's six basic emotions. It was created by Houdna-khilouf and last updated on HuggingFace in April 2026.

Use Cases

• Train emotion classification models based on Ekman's six basic emotion labels.
• Benchmark NLP model performance on Algerian Arabic (Darija) text.
• Study cross-cultural emotion expression based on social media comments from YouTube, Facebook, and Instagram.
• Develop dialect-specific sentiment analysis tools for Algerian Arabic.

Strengths

• 6,000 manually annotated data points provide a substantial corpus.
• It is described as the first large-scale dataset for emotion detection in Algerian Arabic dialect.
• Annotations follow the established Ekman's six basic emotions framework.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but other structural details like file formats and license are unknown.
• Data may reflect geographic and platform bias inherent to its social media sources.

Provenance

Source

Houdna-khilouf on HuggingFace.

Collection Method

Manually annotated comments collected from YouTube, Facebook, and Instagram.

Freshness

Last updated 2026-04-17 12:05:18.

Geography

Algerian Arabic dialect (Darija).