lucyd is a music recommendation tool developed by four graduate students in UC Berkeley's Master of Information and Data Science (MIDS) program.
lucyd seeks to demystify the typical "black box" music recommendation algorithm. When you log in to Spotify or Apple Music, you can only assume how these providers develop your weekly recommended playlist.
lucyd offers something different. Our algorithm focuses on you--how do you perceive a song, and how can that be used to find new ones?
lucyd is not for someone who just flips on the radio and rocks along to today's Top 40 hits. It is a tool for music "Savants and Enthusiasts"--the people whose favorite pastime is perusing music blogs or the depths of their streaming service. lucyd is for users who are willing to put in a little effort to find the next song to hit their "Most Played" list.
Designing lucyd required several steps:
- Algorithms: Outline framework and requirements of a music recommendation algorithm based upon user-determined "tags", rather than song and/or artist attributes.
- Data Source(s): Identify potential data sources to "teach" the recommendation algorithm.
- Visualizations: Develop visualizations that help the user understand the songs in their playlist, and the ones nearby
- Application Framework: Combine all of these elements into a live web application that allows the user to influence and learn from the songs recommended by lucyd.
Product Design & Cloud Architecture
github | LinkedIn
About Tim
Visualizations & Project Management
github | LinkedIn
Hanna is a technology consultant in the Customer Experience (CX) industry. She is currently based in Bozeman, Montana, providing solutions and expertise to large enterprise clients as they design, maintain and improve their CX programs. Keeping with her interest in CX and data visualization, Hanna was responsible for the user experience and interactive infographics in lucyd's user interface.
Primary Skills Developed: HTML, CSS, JavaScript & D3
We developed lucyd during the course of our Capstone semester in UC Berkeley's Master of Information and Data Science (MIDS) program. MIDS offers a comprehensive curriculum in data science, including courses focused on computer science, machine learning at scale and legal, policy and ethical issues.