This paper examines the musical boundaries that emerge from the distinct consumption patterns of rap audiences. Using the actual listening histories of around 1000 French users of the music streaming platform Deezer, we apply dimensionality reduction and clustering methods to explore the musical boundaries that emerge from distinctive audience consumption patterns, with a particular focus on rap music. We show that these boundaries exhibit salient thematic distinctions, and each region of the map holds its own combination of themes. Focusing on six demographic groups based on age and gender, we find that each exhibits a unique pattern of music preference across the highlighted boundaries. Our findings deepen and renew our understanding of the dynamics in which music boundaries are formed, and highlight the importance of studying and comprehending these dynamics, showcasing one way to shed light on that matter.

中文翻译：

---

研究流媒体平台时代的音乐分类法：通过实际消费数据从说唱音乐中获得见解

本文研究了说唱观众不同的消费模式中出现的音乐界限。利用音乐流媒体平台 Deezer 的大约 1000 名法国用户的实际收听历史，我们应用降维和聚类方法来探索不同观众消费模式中出现的音乐界限，特别关注说唱音乐。我们表明，这些边界表现出明显的主题差异，并且地图的每个区域都有自己的主题组合。专注于基于年龄和性别的六个人口群体，我们发现每个群体在突出显示的边界上都表现出独特的音乐偏好模式。我们的研究结果加深并更新了我们对音乐边界形成的动态的理解，并强调了研究和理解这些动态的重要性，展示了阐明这个问题的一种方式。