Exploit musical data to demonstrate machine learning techniques, mainly classification and clustering. The final product is a web application that allows users to provide some information about their music preferences and get recommendations based on the data.

• Python
• Django
• NumPy
• Pandas
• Scikit-learn
• Spotipy

Install the dependencies using the following command:

pip install -r requirements.txt

To run the web application, use the following command:

python manage.py runserver

• Use user's data (authentication required)

• Use a public playlist

• Visualization of musical data

  • Provide multiple visualization options for user's musical data
  • Provide multiple other summarization options (artists, playlist, etc.)

• Recommendation system

  • Provide multiple recommendation options (mood, genre, etc.)
  • Constraint-based recommendation (known artists, speed, etc.)
  • Prompt-base recommendation (LLM)

• Playlist generation

  • Provide multiple playlist generation options (mood, genre, etc.)
  • Constraint-based playlist generation (known artists, speed, etc.)
  • Prompt-base playlist generation (LLM)

• Query musical data from Spotify API

• Clustering of musical data

• Classification of musical data

• PCA and t-SNE for dimensionality reduction

• Extract musical features from raw audio files

• Fourier transform

• Extract linguistic features from lyrics

• Django web application