Smart Objects and Technologies for Social Good. Third International Conference, GOODTECHS 2017, Pisa, Italy, November 29-30, 2017, Proceedings

Research Article

The Fractal Dimension of Music: Geography, Popularity and Sentiment Analysis

  • @INPROCEEDINGS{10.1007/978-3-319-76111-4_19,
        author={Laura Pollacci and Riccardo Guidotti and Giulio Rossetti and Fosca Giannotti and Dino Pedreschi},
        title={The Fractal Dimension of Music: Geography, Popularity and Sentiment Analysis},
        proceedings={Smart Objects and Technologies for Social Good. Third International Conference, GOODTECHS 2017, Pisa, Italy, November 29-30, 2017, Proceedings},
        proceedings_a={GOODTECHS},
        year={2018},
        month={3},
        keywords={Music data analytics Hierarchical clustering Sentiment pattern discovery Multi-source analytics},
        doi={10.1007/978-3-319-76111-4_19}
    }
    
  • Laura Pollacci
    Riccardo Guidotti
    Giulio Rossetti
    Fosca Giannotti
    Dino Pedreschi
    Year: 2018
    The Fractal Dimension of Music: Geography, Popularity and Sentiment Analysis
    GOODTECHS
    Springer
    DOI: 10.1007/978-3-319-76111-4_19
Laura Pollacci1,*, Riccardo Guidotti2,*, Giulio Rossetti1,*, Fosca Giannotti2,*, Dino Pedreschi1,*
  • 1: University of Pisa
  • 2: ISTI-CNR
*Contact email: laura.pollacci@di.unipi.it, riccardo.guidotti@isti.cnr.it, giulio.rossetti@di.unipi.it, fosca.giannotti@isti.cnr.it, dino.pedreschi@unipi.it

Abstract

Nowadays there is a growing standardization of musical contents. Our finding comes out from a cross-service multi-level dataset analysis where we study how geography affects the music production. The investigation presented in this paper highlights the existence of a “” musical structure that relates the technical characteristics of the music produced at regional, national and world level. Moreover, a similar structure emerges also when we analyze the musicians’ popularity and the polarity of their songs defined as the mood that they are able to convey. Furthermore, the clusters identified are markedly distinct one from another with respect to popularity and sentiment.