The Angry Tweets dataset is a collection of anonymized Danish Twitter data that has been annotated for sentiment analysis through crowd-sourcing. Here are some key details about the dataset:

• Tasks: The dataset is suitable for text classification, specifically sentiment classification.
• Languages: The dataset is in Danish.
• Size Categories: The dataset falls in the size category of 1K<n<10K.
• Annotations Creators: The annotations were created through crowdsourcing.

Each entry in the dataset has a tweet and an associated label. The label can be "positiv", "neutral", or "negativ" for positive, neutral, and negative sentiment, respectively. The dataset is split into a training set and a test set, with the test set being 30% of the dataset, randomly sampled in a stratified fashion.