Master's Thesis

Conception and Implementation of an Interactive Data Visualization for Researching Musical Reconstructions

This master's thesis explores the conception and implementation of an interactive data visualization designed for researching musical reconstructions. The study covers the fields of musical reconstruction and user interface design while also analyzing the background processes of the application in relation to Spotify API datasets. The work is particularly relevant for developers and data scientists interested in non-linear representations of audio-specific datasets. Additionally, it provides value for music enthusiasts and media professionals looking to conduct specialized research within Spotify’s music databases.

The foundational part of this thesis discusses the characteristics of musical content in the context of Music Emotion Recognition and Music Information Retrieval. It further examines the parameters derived from the findings of the Spotify Research Team and defines the concept of musical reconstruction based on different forms of reproduced music. The study also outlines key aspects of data visualization, comparing them to the requirements of the developed interactive system.

The conceptual phase focuses on selecting appropriate chart types and mapping datasets while considering aspects such as clustering and ranking representation. The implementation process investigates suitable development environments and discusses technical processes, including the search algorithm for detecting musical reconstructions. Finally, the thesis consolidates its findings, presents a solution approach, and offers an outlook on future projects and research opportunities.

Below, you will find a demo video showcasing the developed application build with Unity.

Yannick_Schmitz_Mastertarbeit.pdf

PDF herunterladen • 6.04MB