The goal of this project is to better understand the measurable characteristics of musicals, and the patterns that emerge among different types of shows.
This project currently consists of three modules:
crawl: code for scraping lyrics and accessing APIsmunge: scripts for formatting and joining dataanalysis: R code for data exploration
- "Will West" [email protected]
For this project, several types of data are needed. First, lyrics are needed to capture the topics and lexical characteristics of each song and show. Metadata about each show and song is also necessary, so we can model the patterns that emerge between different composers and artists.
While there has been a lot of work done in analyzing music in general, often musicals are lumped into the soundtrack category without specific attention given to their unique characteristics. This results in recommendation systems and music discovery applications that don't properly capture a user's tastes when it comes to musicals.
Musicals tend to delve heavily into one or more topics that may set it apart from other musicals. While lyrics in traditional genres may address specific emotions (love, anger, etc.) or cultures (pop culture, politics, etc.), a musical can revolve completely around a single topic. Take for example the popular broadway musical Next to Normal. The entire musical is focused on mental health, treatments, and the way both can positively and negatively impact a family.
While traditionally albums are composed and performed by a single artist, musicals often are composed by a single composer and performed by several different artists. A listener's opinion of a musical album may be impacted by both an affinity towards the composer/lyricist, and a mixture of preferences for each artist featured on the album.