This project aims to investigate the complexities of music genres and features that present in songs. A song can overlap between many different genres, making it difficult to formally classify it. The dataset used contains audio features associated with Spotify tracks. We selected tracks from the 10 most popular and 10 least popular genres to get a wide range of song selection. After cleaning the data, we ran several different clustering models to create visualizations. Our results show that certain features had a greater impact on clustering than others. We measured this by finding the mean and median values of each feature for each of the clusters.
Repository navigation
forked from COGS118B/group_template
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|