Dataset of tweets in English language about the COVID-19 pandemic for binary sentiment analysis

This dataset is aimed to the task of sentiment analysis in tweets about the COVID-19 pandemic. There are 3 versions of the dataset, composed by 186,000, 132,000, and 82,000 tweets in English language with stopwords removal, respectively. Positive tweets have polarity equal to 1, while negative tweets have polarity equal to 0 in all versions. All datasets were selected, cleaned and organized from the public dataset available at . The datasets are accompanied by embedding matrices generated from the pre-trained Word2Vec shallow neural network available at .