Inspiration
The drive to build something unique and cool
What it does
It receives haikus and determines the sentiment value of each line by processing the 5-7-5 syllable format. It then assigns a unique musical accompaniment to each line which would be played back along with the haiku. It chooses from approximately 343000 different tunes that were all produced by a member of our team.
How I built it
We used the Twitter Streaming API to receive a stream of tweets that were decoded from their JSON format. We then used an algorithm to detect the 5-7-5 syllable format of haikus. The Indico Machine Learning API determined the sentiment value of each line and we assigned it to one of 5 possible sentiment values. After that, we used another algorithm to randomly match it up to a list of 70 musical phrases to create different combinations and therefore unique sounds each time. Finally, the device plays back the user's name followed by the haiku with the accompanying music.
Challenges I ran into
Understanding and decoding the JSON format in which the tweets were received
Accomplishments that I'm proud of
Working together with a team of students with different skill sets to create a technically challenging task.
What I learned
Learned about API and machine learning and also Python programming language.
What's next for BirdSongBot
The unique song produced can be played backed on the sender's phone rather than the computer that was running the code.
Log in or sign up for Devpost to join the conversation.