Cohort 19 Capstone Project for the Graduate Certificate of Data Science at Georgetown University School of Continuing Studies.
Adam Goldstein (Project Coordinator) - https://www.linkedin.com/in/adammgoldstein1/
Patricia Merino - merinopc@gmail.com | www.linkedin.com/in/patriciamerino7
Navneet Sandhu - nav.sandhu03@gmail.com | www.linkedin.com/in/navneetksandhu2
Nicholas Merkling (Statistician) - merknc08@gmail.com | https://www.linkedin.com/in/nicholas-merkling/
We believe that creating playlists driven by lyrical content can give the user a glimpse into thematic sequences that exist within their “liked” tracks.
We believe sonic variation in a playlist can be achieved by ordering the tracks according to a valence-energy curve, thus delivering a captivating listening experience.
Spotify offers a plethora of playlists ranging from types of moods to genres. However, some playlists, such as “POLLEN”, declare itself as a genre-less playlist. Our playlist methodology expands on the idea of a genre-less playlist by connecting tracks through their lyrical message and themes rather than genre. Our product is called a Thematic Sequence Playlist (TSP). The TSP is created by collaborating with a Spotify user’s “liked tracks” to create a ten to twenty track playlist where tracks are linked through a structured lyrical theme.
POLLEN is “a playlist based around a radical premise: It was not organized by genre.” This playlist has 410,000 followers and is made up of 120 tracks. A testimony to Pollen’s listeners from Hope Tala, a producer of an artist with 450,000 monthly listeners, exclaims “We’ve been on other playlists that have higher follower numbers, but the engagement on Pollen is mad….This is not like a mid-day cafe background playlist; this is a, I-need-to-listen-to-this-today” (1).
Personal Music Collection Database - Train model to find features of “importance” to learn in order to generate TSPs.
Spotify API - Use liked tracks to make TSPs based on trained model from Personal Music Collection Database
Kaggle Data File - Train model on lyrics