Classical music orchestras are vital to the cultural scenes of both Germany and USA. Despite ongoing discussions on musical canon, gender equality, and repertoire innovation, empirical studies on the actual frequency of performances of individual classical music works in both countries are scarce. In this study, concert programs of professional orchestras from the 2019/20 and 2023/24 seasons were collected via web scraping and enriched with metadata from the various online sources using data linkage. In addition to a detailed descriptive statistical analysis, we determined key factors of stage performance frequency using random forest models. Based on these factors, canonical repertoire types were then identified using Latent Class Analysis. Internal factors for stage success of individual works from these repertoires were subsequently determined using Mixture Regression. Results suggest that normative criteria tend to play a more decisive role for program selection in Germany, while the USA lean more towards popular criteria. Four repertoire types are distinguished, primarily based on work and artist prestige, work age and work duration. Generally, programming in the US appears to be more diverse and innovative, and a pre- and post-pandemic comparison of the combined program data from the two countries reveals little differences. These empirical results support and expand on previous musicological studies of canon and can serve as orientation knowledge for cultural policy. Furthermore, the available dataset can also be used by future studies to examine the ways of curating classical concert events and for more fine-grained sub-repertoire analyses.

中文翻译：

---

又是巴赫、贝多芬和勃拉姆斯？对 21 世纪初德国和美国古典管弦乐事实上经典的计算观点

古典管弦乐队对德国和美国的文化舞台都至关重要。尽管关于音乐经典、性别平等和曲目创新的讨论一直在进行，但关于两国个别古典音乐作品的实际演出频率的实证研究却很少。在这项研究中，通过网络抓取收集了 2019/20 和 2023/24 乐季专业管弦乐队的音乐会节目，并使用数据链接丰富了来自各种在线来源的元数据。除了详细的描述性统计分析外，我们还使用随机森林模型确定了阶段表现频率的关键因素。基于这些因素，然后使用 Latent Class Analysis 确定典型库类型。随后使用混合回归确定这些曲目中单个作品舞台成功的内部因素。结果表明，规范性标准在德国的项目选择中往往起着更决定性的作用，而美国则更倾向于流行标准。分为四种曲目类型，主要基于作品和艺术家的声望、工作年龄和工作持续时间。总的来说，美国的项目似乎更加多样化和创新，对两国疫情前后的综合项目数据进行比较，发现差异不大。这些实证结果支持并扩展了以前的经典音乐学研究，可以作为文化政策的定向知识。此外，可用的数据集也可以用于未来的研究，以检查策划古典音乐会活动的方法，并进行更细粒度的子曲目分析。