Analyzing the Relationship Between Spotify Audio Features and Streaming Activity Across Platforms
In some free time, I wanted to explore if there were relationships between Spotify's audio features and streaming activity across various platforms. I pulled data from Spotify's API and found a public dataset containing streaming metrics for approx. 4500 songs.
The goal of this project was to use Python to see if certain audio features could be clustered into certain genre/mood song types and then to analyze how streaming varies across these clusters. I used Principal Component Analysis and K-Means clustering to see how they could group songs. Turns out there are some trends among which song groups performs better on different platforms!
LINK TO CODE: https://github.com/Amandapoor2000/SpotifyFeature-Streaming_ProjectCode/blob/main/Spotify_Streaming%26AudioData.ipynb
Pitching the Probabilities.
For my honors thesis project, I chose to look at every pitch from the 2021 MLB regular season to see if a certain pitch type (breaking, fastball, or off-speed pitch) was more likely to result in a desirable pitch (from the pitchers perspective). Click to read my honors thesis project!
Predicting Spotify Song Popularity.
In my Data Mining class, I worked with 2 other classmates and did a project using SPSS Modeler and a Spotify Data Set, comprised of over 600k songs defined by different features.
Simpson’s 3/8 Rule.
In my Numerical Analysis class, I worked with classmate Allie Texter on a report on Simpson’s 3/8 rule, a calculus approach to solving for the area under a curve. We used MatLab to assist with the calculations and to evaluate whether other methods were more accurate in finding area approximations.