An exploratory analysis of Spotify song tracks to predict song popularity using various Machine Learning Models.
Spotify is an online audio streaming and media service provider application. There are over 450 million monthly active users and it is famous application among the users all over the world. The Spotify application provides the user with various features and options. One such feature is their song recommendation system, It is very famous for this particular feature.
Spotify's most impressive piece of engineering is its use of convolutional neural networks (CNN). Using CNN, Spotify analyses raw audio data such as the song's BPM, musical key, loudness, etc., to classify songs based on music type and further optimise its recommendation engine.
We have done an exploratory analysis of Spotify song tracks to predict song popularity using various Machines Learning models.
The data is about song and its features that are used to predict the popularity. There are 16 columns. 13 of which are song attributes. Here are the 13 track attributes: acousticness, danceability, durationms, energy, instrumentalness, key, liveness, loudness, mode, speechiness, tempo, timesignature, valence.
- Setting up the environment
- Importing libraries
- Uploading the data
- Data Cleaning
- EDA
- Feature Engineering
- Model Creation
- Predictions