Multimodally visualizing your Spotify music tastes
Building a 3D Spotify Data Visualization Website + a version in augmented reality
2021
Musicscape.us
Tools
Three.js, Javascript, HTML/CS, React
Live website
As an avid Spotify listener myself, I was excited when I found how accessible and in-depth Spotify’s API was. Using this, I created a website that visualized your music tastes by abstract qualities such as valence (mood), acousticness, and energy. Each song is a single dot that’s plotted on a three-dimensional graph according to these qualities.
This video below uses audio. Please unmute! 🎶️
2022
Musicscape for Augmented Reality / Hololens 2
Time
2 weeks, solo project
Tools
Unity, C#, prototyping, data visualization, Figma, MRTK
In this project, I brought my Musicscape website to an AR space. Musicscape for AR is a mixed-reality app that visualizes your music tastes on Spotify by abstract qualities such as mood, acousticness, and energy. Each song is a single dot that’s plotted on a three-dimensional graph according to these qualities.
I learned how to use the Mixed Reality Toolkit and how to develop for the HoloLens. I experimented with the recognition and usability of different gestures. Additionally, I explored how people navigate data in a space, as well as how to make compelling microinteractions. Please reach out if you’d like to talk about this in-depth.
This video has audio! 🎶️ (Song from artist Monster Rally)
2021: Spotify Profiles
Tools
This was the final project for my computer science class. Usually, artists are labeled with a certain genre. But many artists have a wide range of styles that expand across multiple genres and styles. With this app, I wanted to visualize the different styles of an artist and how their songs relate to each other.
In the video demo below, you can see the different moods of Lil Uzi Vert as he expands on the domain of rap ranging from his emo-sounding to his dance-oriented music.
Some challenges that came with making this project was how to see the eleven properties such as valence, mood, danceability, and loudness of music all at once on a 2d screen. I approached this with algorithms such as k-clustering and dimensional reduction to group similar songs together through similar characteristics across those eleven properties. While I got an A for this project, I was not satisfied with the visuals due to the limited time. I may improve the visuals in the future.
This video below uses audio. Please unmute! 🎶️
