“Would It Play in Peoria?” Patterns of Mass Culture in Progressive America

During the late nineteenth and early twentieth centuries, a massive expansion of popular entertainment reshaped American society. Once viewed with distrust, touring performers became increasingly central to public life throughout the United States, knitting together a myriad of distinct communities into a national mainstream. This process was dominated by a handful of New York-based firms, theatrical syndicates that connected tens of thousands of artists to a weekly audience of millions. Seeking greater profits, these syndicates sought information about the tastes and preferences of their far-flung customers, data that flowed from local theaters to corporate headquarters before being returned through booking and advertising. Repeated over decades, the result was a revolutionary form of commercial culture uniquely attuned to the national audience that it had helped construct.

Scholars have long pointed to the importance of this burgeoning industry, linking it to the changing racial dynamics, gender relations, and structures of consumption that defined the Progressive era. However, research on the topic has long struggled with the sheer size of the phenomenon under description. Vaudeville—the most popular of the period’s theatrical forms—comprised between ten and fifteen thousand performers playing in over a thousand theaters, forty-plus weeks a year for three decades. Despite the ready availability of evidence documenting its evolution, researchers have lacked the ability to analyze trends at scale. As a result, it has been difficult to examine the development of American mass entertainment as it was enacted on the ground—to find out what, as the expression went, would “Play in Peoria.”

Developed in collaboration with computational linguist Tom Lippincott, this project employs new forms of data analysis to recreate the touring patterns of American Vaudeville, tracking the genre as it evolved through the interaction between local demand and centralized decision-making. Using text-mining techniques to create a database from the touring schedules published in the trade-press Variety, the project analyzes these records with “Starcoder,” an unsupervised machine learning model designed for rich network data. The result is a database mapping the movements of thousands of touring performers between cities throughout the United States, creating a detailed picture of the first era of American mass entertainment

Initial results—demonstrated in the maps and graphs that form the poster’s visual centerpiece—provide insight into the complexity of the relationship between local and national culture during the early twentieth century. Large-scale trends became increasingly prevalent as a distinct mainstream was established. At the same time, regional tastes not only survived but, in certain cases, seemed to strengthen, as traveling entertainers offered audiences a vehicle for the articulation of identity. Equally important, the rich data produced by the project opens up a host of new avenues for potential historical research, enabling new questions about topics such as the dynamics of race in mass culture or region in entertainment consumption. Combining cultural analysis with machine learning to both produce and examine large-scale historical data sets, this project demonstrates the potential—and lays out a roadmap—for similar interdisciplinary work in the future.