Emotions and Irony per Gender in Facebook

This paper describes a Spanish dataset collected from Facebook that has been labelled with emotions, irony and author’s gender. The inter-annotator agreement shows the difficulty and high level of subjectivity of the annotation task, especially with respect to irony. Statistics of the corpus show the relationship among topics, emotions, irony and author’s gender. For instance, females used more emotions than males (mainly positive emotions), males were more ironic than females (at least in this dataset), and politics is the topic addressed more ironically and with more negative emotions. A social analysis of the results goes beyond the scope of this paper but being the comments mainly about (some of) the Spanish politicians, we are not surprised of the results. The dataset is publicly available for research and social analysis purposes with the name EmIroGeFB at http://ow.ly/uQWEs.