HaSpeeDe 2018 dataset

The HaSpeeDe dataset collects 8000 Facebook comments and tweets annotated for the presence of hate speech. The dataset has been used in the context of the HaSpeeDe task, organized as part of the EVALITA 2018 evaluation campaign (http://www.evalita.it/2018). In order to meet the GDPR requirements, texts have been pseudonymized replacing all original IDs in both datasets with newly-generated ones. Mentions, emails, person names (excluded public person names), and phone numbers have been masked with, respectively, the labels MENTION, EMAIL, PERSON, PHONE, followed by a number to distinguish between different entities of the same kind within the same text.