Playlist creation system based on audio content and contextual information of musical pieces

In the context of this thesis, a system – RESTful (Representational State Transfer) WEB Application Programming Interface (API) has been created which implements the theoretical principles of music track suggestions.

The playlist API is accessible through public URLs, is based on algorithms of graph theory and data mining and suggests music tracks based on queries from a consumer of the playlist API. Its purpose is to operate as an integrated microservice so that anyone can consume it independently of their platform, simply by taking advantage of the HTTP communication protocol with the playlist API.

The playlist API is built with modern technologies and relies heavily on services offered from companies that dominate the online music consumption (e.g. Spotify).

This thesis examines the most common practices of algorithmic music track suggestions as well as the theoretical bases of the playlist API itself. It analyzes the design, analysis and methodology of the technical part of the implementation and presents its functionality as well as its results.