Project Title:
- MuJi
By:
- Anna Yuen
- Naoki Okada
- Titapa Chaiyakiturajai
- Manish Aryal
Problem Statement (i.e., what is the problem?)
- Music has great impact on our emotions. It can help uplift us when we are down, or get us motivated when feeling lazy. However, it's a hassle and time-consuming to create new playlists for different moods. It's even more time-consuming to search for an appropriate playlist that matches your mood.
How do you solve the problem?
- Our solution involves an emotion detector that detects the user's current emotion (through audio--what the user says) and matches that to an appropriate playlist. Now, the user can enjoy a good playlist at any time, in any mood, effortlessly!
List of all the features that your team will implement. Your team will be held accountable for the featurs you list!
- Recognizes emotions through audio--what the user says
- Automatically picks songs according to your emotion and plays them
- User will have the option of saving the played song into a certain playlist
What data will your prototype be using and collecting?
- We'll be using the data from the user (their response to "How was your day?")
- We'll also be using songs from Spotify API with extra information that represents the properties of the song (i.e. dancability, energy, liveness, etc.)
Any algorithms or special techniques that will be necessary.
- We'll be using sentiment analysis through Google's API to determine the emotion based on user's response (their audio response).
- We will provide user with all universal emotions and their corresponding calculated percentage of how much of that emotion is present.
- We'll use the data we get from user's response and create our algorithm to present the most relevant songs to user's emotion. Their song selection will be taken into account (whether they choose to skip the song or not).
Electronic Mockups:
- Way cool! Use of speech recognition is fantastic and bold. Total yes.