The matter of taste Genre compatibility in consumer profiles Mapping listening behaviour patterns to identify genre compatibility in the profiles of the Last fm users.

In a world where every type of music is available from every online medium, listeners can be overwhelmed by the amount of content. Several online platforms and music providers offer a recommendation system and a personal account for their users, acting as digital cultural intermediaries by recommending songs while mapping the users’ listening behaviour. One of these online platforms is Last fm, a website that connects different music streaming services into one user account, in which they recommend music and analyse the listening behaviour for an improved experience for the user, and provide live statistics on listening behaviour. This vast amount of music listening behaviour data of Last fm is anonymized into a dataset made available for non-commercial research. In this thesis, the Last fm database is used to map consumption patterns and find genre compatibility using cluster analysis and linear regression models. The cluster membership is then linked to the user data consisting of age, gender, country, listening intensity and diversity score. Five clusters of Listening behavioural patterns are found and are compared. The way different genres correlate with each other and with diversity gives insight into the genres compatibility and patterns.