Quantitative analysis of K-Pop writers’ network by using K-Pop lyrics

Abstract

Music is one of the most important aspects of culture that people can easily enjoy. Especially, pop music is influenced by the lifestyle of the people at that time, and lyrics implicitly contain much information on that lifestyle. In this study, we collect and quantitatively analyze Korean-pop (K-pop) lyrics provided by a music streaming service. We find that the K-pop consuming market exhibits a power-law distribution in the number of released songs for each artist. We create an artist-writer collaboration network by using a bipartite network between artists and writers who collaborate on a song and measure the network properties that obey the power laws in this network. All exponents of the power laws are -3, which can be explained by using a preferential attachment, the rich get richer. We found that both Zipf’s law and Heap’s law hold for K-pop lyrics even in different music styles and genres, and we created a style-similarity network among K-pop writers by incorporating the cosine similarity of word frequencies to uncover the relationships between K-pop writers.