Analyzing sentiment in a graphical way, using mixtures of two oil paint. 

Partner: Steven Simon

For the ITP winter show, I wanted to analyze the user's sentiment based on their word usage. People have a lot of interests in what others think and feel on the same issues. Sentiment analysis can be an excellent source of information and can provide insights that can either determine marketing strategy, improve customer service and product messaging.

This project is about watching user and other participant's sentiment visually. 

We have set different words of moods and divided into two groups, positive and negative. When the user selects their word, two oil color paints will drip into the big bowl. After placing a piece of paper, each user can make a unique marble paint based on their moods. They can also see and compare the collected sentiment of the whole participants.


Users need to drag and drop the mood blobs to the centered blob. We created a blob look using Paper.js. So when 2 circle meets each other, it creates the vertexes.  Each blob has a different words.(Positive/Negative) When they drag their mood blob to the bigger circle in the center, the peristaltic pump will be triggered and the screen will reset. 

Artboard 1 copy 7.png

Idea sketch


User test

After several user tests, we made a progress of our ideas. At the very first beginning, we limited people to vote their opinions only once. But participants wanted to have more fun with dragging and dropping the blob multiple times. So we decided to go much simple way. We added the function that automatically resets the small blob position. We allowed users to experiment freely by not limiting the count for choosing the words.

This is the link of the screen.