Data Science

Encoding Music Final Project

This is my portion of a project done in collaboration with Will Betts Cope and Shoki Kawasaki. The overall goal of the project was to analyze the use of parallel fifths and harmonic intervals in Schein's choral works as opposed to the arrangements produced by Bach. The full project will be posted soon if Will and Shoki have not provided it somewhere already. My portion of the project, however, was creating a visual representation for harmonic ngrams in the MusicXML files. I wrote a python script in blender that generates a chart, which I will call a "trace graph." This trace graph maps harmonic intervals between one part and the three others at a given note, creating a triangle. The colors generated for the triangles are also reflective of the coordinates of the vertices on the chart. The current ranking for the intervals is a subjective ranking of dissonance, which shows how Bach's arrangement is more unusual than Schein's. However, the axes for the chart can be modified easily either by inputting an array, or specifying ranges for numbers to input. Using Google's model viewer, I have made eight charts available to view using the buttons above. Lastly, animated trace graphs are also possible in the script I wrote, and go through a series of harmonic intervals, but due to the time constraints of finals, I needed to leave the exploration and presentation of that part of the project for a later date.

Louvain Networks with Spotify's API

This project was done in collaboration with Logan Griffin and Will Betts Cope. The goal of the project was to analyze the various music qualities in Spotify's API through the use of Louvain networks. Our Encoding Music class created playlists, which we then used as a dataset for analysis. Here we have binned qualities such as energy, speechiness, valence, loudness, and liveness to connect songs from our playlists. You can see the interactive network above.

Analysis of Cases and Weapons in Counter-Strike: Global Offensive

This project was a collaboration with James Simbolon for our Data Science Final. We analyzed different qualities of weapons and cases in the Game Counter-Strike: Global Offensive. For the Full Project and report, visit the Github page: https://github.com/Jsimb174387/case-analysis