10. Application of Computational Methods to the Analysis of Topics in the Piano Sonatas of W.A. Mozart

Brandon Zlotnik Baldwin-Wallace College, Rachel Fogle Baldwin-Wallace College, Miranda Lemmer Baldwin-Wallace College
Faculty Sponsor(s): Jessica Narum Baldwin-Wallace College, Andrew Watkins Baldwin-Wallace College
Drawing resources from the musicological writings of Ratner, Monelle, Hatten, Allanbrook, Rumph and others, we have been working to extract and visually represent musical topics in the piano sonatas of W.A. Mozart through the use of computer data science tools. Musical topics are defined as sections of music with common characteristics which signal cultural meaning for the listeners through reference to popular styles such as waltzes and marches. Topics were brought into the focus of musicologists beginning in 1980 with Ratner’s book, Classic Music, and have persisted as a way of accessing and understanding classical music. Scholars have published many works on topics, but most have left one aspect untouched: a visual and interactive representation of the topical data. Computationally analyzing and visualizing topics could lead to the development of classroom learning tools and software that automatically extracts topical data. Utilizing music21 we were able to abstract a list of significant features including average melodic interval, frequency of chromaticism, and average notes per second, to develop a predictive model. Additionally, we have been analyzing various ways of representing the extracted features such as utilizing the D3.js library, Tableau, and Excel. As more data is gathered, our ability to accurately predict a topic in a given measure increases too. Our most recent research has focused on the idea that a measure could be made up of multiple topics, allowing our accuracy and overall analysis to improve.
Music
Poster Presentation

When & Where

Irwin Library 2nd Floor