The platform lets you search a corpus of public posts from seven social networks, in 30 languages and 20 topic categories, each labelled automatically on a five-level polarity scale. Searches can be filtered by topic, date, language, polarity, category and network, and the results exported as CSV, JSON or PDF. The polarity, category and language annotations are released under CC BY 4.0.
[ Open the ExtremeSentiLex platform ]
If you use it, please cite: S. Pais, I. Tanoli, M. Albardeiro and J. Cordeiro, “Unsupervised Approach to Detect Extreme Sentiments on Social Networks”, ASONAM 2020, pp. 651–658. doi:10.1109/ASONAM49781.2020.9381420
University of Beira Interior
Department of Computer Science
Rua Marquês d'Ávila e Bolama
6201-001 Covilhã
Portugal
+351 275 242 081