This Song Recommendation System allows users to find similar songs based on musical attributes such as acousticness, energy, valence, and danceability. Using Euclidean distances and a graph-based approach, the system suggests songs that share similar characteristics.
- Loads and cleans a Spotify dataset.
- Samples 1,000 songs randomly for analysis.
- Saves the sampled songs to a file (
sampled_songs.csv) for user reference. - Computes song similarity using key musical features.
- Builds a graph-based recommendation system using NetworkX.
- Interactive user input for easy song lookup.
The Euclidean Distance is used to measure the similarity between songs. It calculates the straight-line distance between two points in a multi-dimensional space. In this system, each song is represented as a point in a 4-dimensional space defined by the features:
- Acousticness 🎵
- Energy ⚡
- Valence 😀
- Danceability 💃
The formula for Euclidean Distance between two songs A and B is:
[ d(A, B) = \sqrt{(A_1 - B_1)^2 + (A_2 - B_2)^2 + ... + (A_n - B_n)^2} ]
where each A_n and B_n represents a feature value of a song.
A lower Euclidean distance means the songs are more similar.
- Run the script (
python script_name.py). - Check the file
sampled_songs.csvto see available songs. - Enter a song title and artist name when prompted.
- Receive recommendations for similar songs.
- Repeat or exit the recommendation loop.
- A CSV file
sampled_songs.csvis created containing 1,000 sampled songs. - The system prints song recommendations based on user input.
- Python
- Pandas (for data manipulation)
- NumPy (for numerical computations)
- NetworkX (for graph-based recommendations)
- Scikit-learn (for Euclidean distance calculation)
- The system only recommends songs from the sampled dataset (
sampled_songs.csv). - If a song is not found, check the file and pick a song from the list.
- The threshold for similarity can be adjusted in the script.