1. Twitter Data set for Arabic Sentiment Analysis: This problem of Sentiment Analysis (SA) has been studied well on the English language but not Arabic one. Two main approaches have been devised: corpus-based and lexicon-based.
2. Sports articles for objectivity analysis: 1000 sports articles were labeled using Amazon Mechanical Turk as objective or subjective. The raw texts, extracted features, and the URLs from which the articles were retrieved are provided.
3. Multimodal Damage Identification for Humanitarian Computing: 5879 captioned images (image and text) from social media related to damage during natural disasters/wars, and belong to 6 classes: Fires, Floods, Natural landscape, Infrastructural, Human, Non-damage.
4. Turkish Spam V01: The TurkishSpam data set contains spam and normal emails written in Turkish.
5. A study of Asian Religious and Biblical Texts: Mainly from Project Gutenberg, we combine Upanishads, Yoga Sutras, Buddha Sutras, Tao Te Ching and Book of Wisdom, Book of Proverbs, Book of Ecclesiastes and Book of Ecclesiasticus
6. Gender by Name: This dataset attributes first names to genders, giving counts and probabilities. It combines open-source government data from the US, UK, Canada, and Australia.
7. Wisesight Sentiment Corpus: Social media messages in Thai language with sentiment label (positive, neutral, negative, question).