Two corpora of Twitter messages in Spanish and Catalan categorized into three stance classes: AGAINST, FAVOR, and NEUTRAL. The collection captures social media discourse from Barcelona and Terrassa during specific political windows in September 2018 and February-March 2019.
Use Cases
- Train a stance detection model using the tweet text and the three-class stance labels.
- Perform cross-lingual sentiment analysis to compare political discourse patterns in Spanish versus Catalan.
- Analyze geographic variations in political stance by comparing the Barcelona and Terrassa subsets.
Strengths
- Bilingual text data provided in both Spanish and Catalan.
- Annotated with three distinct stance labels: AGAINST, FAVOR, and NEUTRAL.
- Data collected from two specific geographic locations: Barcelona and Terrassa.
- Temporal coverage spanning 12 days in early 2019 and a period in September 2018.