APPLYING DATA MINING AND DATA VISUALIZATION WITHIN THE SCOPE OF AUDIO DATA USING SPOTIFY

The aim of this research is to put forward an overview of applying data mining and data visualization within the scope of audio data from a dataset of Spotify. The research starts by presenting background information of these two fields and their influence on music industry. The paper includes explanation of the most essential concepts and their role. The research is concentrated on analysis of audio features of the tracks of Spotify’s Top Songs in 2017 playlist and tries to highlight the common patterns behind the audio features of these songs. For the purposes of this research, Spotify datasets are used as practical scenario. For this reason, more detailed information is given about songs features, what are they, what do these top songs have in common and why do people like them. The result of the study showcase how singers and song makers can leverage the power of data visualization and data mining to help trying to predict one audio feature based on the others, look for patterns in the audio features of the songs and see which features correlate the most. to study the relation between learning and preference. They showed that passive exposure to melodies built in an entirely new musical system led to learning and generalization, as well as increased preference for repeated melodies.