Enhancing Spotify Experience with Data-Driven Recommendations and Visualizations
-
© Baybali
- Studiengang: Data Science (Master)
- Modul: Projekt Data Science
- Zeitraum: Wintersemester 2023/2024
- Prüfer: Prof. Dr. Anika Groß
Musicians and DJs aim to discover unique musical pieces and blend them into distinct performances. There are numerous methods for discovering new music, ranging from vinyl collecting to using digital music platforms recommendation algorithms. The Spotify Application, one of the most popular and widely used ways to listen to music, has always been fueled by discussions with fellow musicians about how to best use Spotify as a musician. When I discovered the Spotify API, the focus of the Winter Semester 23/24 Data Science Project became clear: optimize Spotify usage to save time, effort, and amplify creativity for personalized music discovery for music enthusiasts who listen and perform. This project focuses on creating unique playlists and analyzing the features of its tracks using Spotify's API, Python, Flask, and Tableau Visualization tools. The Users can handpick their favorite artists, which allows them to create playlists that are tailored to their preferences. Users can use Tableau integration to visually understand the composition of their playlists and access additional information for each track, such as the official download link. Beyond playlist curation and analysis, the project aims to use advanced analytics techniques such as improved K-Means clustering with UMAP and creating live plots. This analysis provides DJs and performers with a thorough understanding of track audio features, allowing them to group songs and save time and effort when preparing a sequential playlist or live performance. All features, including exploratory playlist analysis, track recommendations, and visual dashboards, are wrapped up in a user-friendly web application powered by Flask, providing a smooth experience for musicians and music fans alike.
