Reality Check: Comparing Music Tweets with Music Charts
The ongoing process of digitalization has changed the way music is consumed. Music streaming has become one of the fastest growing sectors in the music business, which makes it a highly relevant source for analysing the listening behaviour of users. Although the data of music streaming services is not publicly available, many services allow users to automatically share the song currently playing on social media, for instance in the form of tweets.
We facilitate these tweets, so called music tweets, in order to monitor the listening behaviour of users. The objective behind this is to analyse whether music tweets can be used to predict the charts. The main part of this thesis consists of three different analyses looking into (i) the temporal relation of music tweets and charts, (ii) the overall agreement of music tweets and charts, and (iii) the genre distribution in music tweets.
The analyses suggests that music tweets react in general faster to change than the Billboard Hot 100. We found, however, that the music tweets do not agree very well with the Billboard Hot 100. Finally, the third analysis showed that the genre distribution in music tweets is substantially different from the genre distribution in music streams in the U.S. We conclude that music tweets do not allow to predict the Billboard Hot 100.