swe-sentiment-dataset

A sentiment annotated dataset for sentiment analysis in Swedish. The dataset contains 2 368 texts in Swedish, all containing emojis.

The data is organized in a .csv file as follows:

text  |  label  |  all_labels

• The sentiments are either positive, neutral or negative, denoted as pos, neu and neg.
• The texts have been labeled by human annotators, where the majority label has been chosen as label.
• all_labels contains a list of all labels by the annotators.

Data source & more information

The data contains Twitter posts and YouTube comments. The paper below describes the creation of the dataset in more detail.

Citation information

If you use the dataset, please cite the following paper:

Berggren, Lovisa. "Sentiment Analysis for Swedish: The Impact of Emojis on Sentiment Analysis of Swedish Informal Texts." (2023).

Feel free to use the dataset for further research or hobby projects.