EMOTIONVIS automatically detects emotions from what people have writen in large conversations. It allows you to estimate the sentiments (positive to negative), levels of arousal (calm to aroused), core emotions (joy, sadness, etc) and fine-grain feelings (frustrated, sorry, etc).
We have shared this prototype with the world to help in the understanding of emotions fromthe patterns within the words that people writeonline, while presenting the results in a veryvisual way for exploration.
Researchers and professionals are invited to use these detection methods with their own conversational datasets.
TEST the emotion engine classifiers by typing a few sentances into the compose tool.
This allows you to immediately see how the classifiers work on your own example texts.
ANALYZE your brand, community or conversation in the dashboard. This is the main tool prototype that allows you to explore your own dataset conversation.
EXPORT all your data including the 38 different emotional detections for each and every post/document in your dataset/corpus.
CONTRAST competitors or different topics using the compare tool to see when one community is more positive than another. There are two modes for juxtaposing your sources; by variance to one another or in absolute terms.