Music Classification Model

In this report, we demonstrate the various design choices and techniques used in the process of building a music classification model based on data extracted using Spotify’s API.

1. We first describe all pre-processing decisions made (normalization, handling missing data, one-hot encoding, etc.); as these pre-processing techniques are used repeatedly across all models that we build and present, we segment this into its own introductory section.
2. We present the working of building a series of classification models using all the remaining features and evaluating the success of these models.
3. We repeat this process, this time adding a relevant dimensionality reduction technique and extracting key insights by clustering our data in the low dimensional space. Finally, we exhibit any additional and supplementary observations that came out during this exploration.