TweetMood is an interactive New (Social) Media Artwork by Rolf van Gelder.
The artwork tries to determine, real-time, the current mood of the world.
For accomplishing that, it gets the latest tweets from Twitter (more or less at random) and have them classified by an Artificial Intelligence Model (ML5).
The AI model has been trained using 20,000 movie reviews.
Movies with a high rating are apparently ‘good’ (‘positive’); low ratings are ‘bad’ (‘negative’).
The tweets are ‘rated’ by the AI model. It returns a value that ranges from 0 to 999. 0 is very negative, 999 is very positive.
Based on that value, a happy or sad smiley is generated.
The images on the ‘ticker’ bar are the profile images of the ‘tweeter’.
The background color of the ‘ticker’ represents the ‘average mood’.
Technology
TweetMood has been written using P5JS.
It uses the CodeBird library for fetching the tweets and ML5 for the Artificial Intelligence model.