SentimentAnalysisParser

SentimentAnalysisParser copied to clipboard

Open to discussion

https://www.media.mit.edu/posts/what-can-we-learn-from-emojis/

this would be great. It would seem that emoji would be a key way to develop sentiment, even more direct than indirect inference from text. The first step would be to do some crawlers, perhaps with Sparkler to figure out a sample distribution of emojis that are widely available. Any takers? :)

I see what you mean. We could also use same approach as this research to find most common emoji in a Twitter dataset, but I agree crawling on wider range of site could give a holistic view of emoji usage, seems harder though :)