I will hold a one-day practical workshop on “Emotions in 50 Years of Pop Song Lyrics: A Text Mining Approach” at the 7th Winter School Fact and Method: Data, Borders and Interpretation in Tartu – Estonia, the 7th of February 2018 (this blog post can give an idea of what we will do). The participation for PhD students is free of charge and, according to the organisers, in some cases, it is possible to reimburse the accommodation. See below a short description of the workshop and some suggested readings.
[University of Tartu: image from flickr]
Emotions in 50 Years of Pop Song Lyrics: A Text Mining Approach
Recent progress in statistical techniques and data mining tools, coupled with the increase in computational power and storage, are allowing human and social scientists to access and examine unprecedented amounts of data. Similarly to ‘natural experiments’ in psychology or behavioural economics, these datasets record naturally occurring behaviours and can be used to track and analyse cultural and stylistic changes.
In this workshop, we will focus on a medium-sized dataset, including 50 years (from 1965 to 2015) of the Billboard Hot 100 songs. The dataset comprises about 5,000 English language pop song lyrics, plus metadata on artists, year, and yearly rank. Our emphasis will be on the analysis of the emotional content of the lyrics (but other aspects can be examined as well if participants in the workshop have any specific interest). Examples of questions we could investigate are: How did the words used to express emotions change in successful pop songs during the last decades? Is there any relationship between the general mood of a song and its success? Are there recognisable trends in pop music, with songs getting happier or sadder? Do artists of different gender express emotion with different words?
The goal of the workshop will be mainly of a practical nature. Participants will learn the basics of text mining using R (https://www.r-project.org) and Rstudio (https://www.rstudio.com), and will be provided with the tools to develop individual projects in the future. Previous knowledge of R is not required, but an interest in learning programming and data analysis is advised.
Dodds, Peter S., and Danforth, Christopher M. (2010) Measuring the Happiness of Large-Scale Written Expression: Songs, Blogs, and Presidents, Journal of Happiness Studies.
Morin, Olivier, and Acerbi, Alberto (2017) Birth of the cool: a two-centuries decline in emotional expression in Anglophone fiction, Cognition and Emotion.
Slige, Julia, and Robinson, David (2017) Text Mining with R. A tidy approach.
Chapter 1: The tidy text format
Chapter 2: Sentiment analysis with tidy data
Online materials introducing R and RStudio: https://www.rstudio.com/online-learning/#R