SENTIPOLC 2014 dataset

The SENTIPOLC 2014 dataset contains 6448 tweets annotated for subjectivity, sentiment polarity, and irony. <p>The dataset has been used in the context of the SENTIPOLC 2014 task (http://www.di.unito.it/~tutreeb/sentipolc-evalita14), organised as part of the EVALITA 2014 evaluation campaign (​​http://www.evalita.it/2014). It is split into a development set (4513 tweets) and a test set (1935 tweets). <p>In order to comply with GDPR privacy rules and Twitter’s policies, the identifiers of tweets and users have been anonymized and replaced by unique identifiers.