Dataset containing tweet and metadata information for multiclass prediction for emotion of tweet.

The data has 4653 tweet records with 5 features and a multiclass target column, referring to the emotion of the tweet.

This dataset is an extension of Cardiff’s tweet_eval dataset, For additional details about the dataset, please refer to the original source: Dataset originally published in “Semeval-2018 task 1: Affect in tweets” by Mohammad et al. (2018):


load_data([data_format, as_train_test, ...])

Load and returns the Tweet Emotion dataset (classification).


Load and return the embeddings of the tweet_emotion dataset calculated by OpenAI.


Load and return a precalculated predictions for the dataset.


Load and return the test data, modified to have under annotated segment.