Spotify's API offers data regarding music features of songs (ex: tempo, popularity and duration) which can be used for modeling and hypothesis testing. Using a sample of 1000 records from the approximately 50,000 recordds, I wanted to:
- Determine if any of the music features of songs can help predict the popularity of a song using linear regression
- Check for a mean popularity difference between three genres (“Anime”, “Hip-Hop” and “Electronic”)
- Build a classification model that can determine if a Spotify is "Anime" genre
- Linear Regression
- Logistic Regression
- ROC Curve
- Analysis of Variance ANOVA
- Hypothesis Testing
- Tukey's Method
- Python
- Jupyter Notebook