Tweet melody is an ambient device that let's the user know about how their close ones are feeling based on their twitter feed. Tweet Melody maps the user's emotion via hashtags on twitter such as #glad, #sad and #mad. These emotion filled tweets are represented as a melody as well as an array of lights which have the emotions color coded as: sad - blue, glad - yellow and mad - red.
From the beginning, the team knew they wanted to work with gauging and translating emotion in some way. The team brainstormed ideas in the space and found that the most compelling way to translate these emotions would be through light. The team initially also wanted to create a caricature of each team member for each board in order to be able to put them all side by side and have them each represent each member tweeting. The second part of this idea was not continued with as it was too obvious and alternates were brainstormed.
The team first spent their time configuring the Neo-Pixel to flash the desired color. Initially the team worked on just making the entire ring light up as the desired color. From there, the team learned how to to code the Neo-Pixel to flash each color five times when that piece of the function was called upon.
The team worked upon understanding how can the Particle get an input or how can a function be called via Twitter and IFTTT. We watched online videos and tutorials to understand the code and functionality. Later, we progressed by trying to integrate the Neo-Pixel and IFTTT.
After the Neo-Pixel was configured the team worked on integrating the Particle with Twitter feed by using IFTTT. The code for Neo-Pixels was broken up into functions that represented three emotions: sad, glad and mad. These functions were then made available online by using the code Spark.function(). The functions were called using IFTTT where the Particle would listen for an event (tweet) which involved one of the following hashtags: #sad, #mad or #glad. Based on the respective hashtag the appropriate function was called.
Additionally, the team worked to make the system alert the user if their loved one had tweeted multiple times in a row with the same hashtag - indicating extreme happiness/distress/anger. Below you can see the response one tweet is sent and when two tweets are sent, respectively.
Different motors and the Piezo speaker were explored as possible ambient ways to map physical and audio cues to emotions.
The team worked with a Solenoid and a transistor to play around with the stroke frequency of the solenoid shaft. The team decided to map different frequencies for each emotion. It was decided to drop the use of Solenoid as the emotions are not coherently conveyed to the user of ambient device.
Servo
The team then decided to adopt the servo to map the different emotions, but then realized that a tune or music would more coherently convey the appropriate emotion.
A Piezo was chosen as the output device along with the Neo-pixel since tunes best convey emotion. The following tunes were used to map the emotions for the prototype.
In a real life ambient device, more subtle audio would have been chosen for each emotion tune.
You can upload files of up to 20MB using this form.
