About this subject

The page shows the distribution of emotions (e.g., joy, anger, sadness, fear) detected in the comments editors leave when they revise a Wikipedia article. The emotion labels are produced by an automated classifier trained on a labelled corpus of revision comments.

It also reports sentiment polarity (positive, negative, neutral) for the same set of comments. The data are derived from the public Wikipedia edit history; each revision comment is extracted, processed, and aggregated over time. Observing shifts in emotion and sentiment helps to understand how collaborative editing evolves, where disagreements arise, and how the community’s mood changes around controversial topics.

Back to the figures