Lecture by Hannah Pivo
Chaired by Naeem Mohaiemen

In the early twentieth-century United States, it became commonplace to make predictions about the future by fitting curves to historical data. Corporate statisticians, university researchers, and federal economists, among others, fit idealized mathematical curves to observed data before projecting them into the future to generate forecasts. Their prognostications informed corporate planning and public policy alike—an early version of what is today often described as ‘data-driven decision-making.’

These early twentieth-century practices relied upon statistical techniques developed in nineteenth-century Europe, especially France and England, where members of statistical societies eagerly (yet cautiously) endorsed the capacity of curves to foretell the future. This lecture examines the early history of using statistical and mathematical curves as forecasting tools and suggests that the delicate confidence placed in these graphic methods relied as much on prevailing practices of geometry and emergent notions of history as it did on the adolescent field of statistics. Through the familiar geometry of curves and the predictable patterns of history, it was thought that data could be made comprehensible, complete, and, ultimately, prophetic.

Fall Thursday Lecture Series events are open to Columbia-affiliated faculty, students, and invited guests. All others interested in attending, please email the SOF/Heyman at [email protected].

Please email [email protected] to request disability accommodations. Advance notice is necessary to arrange for some accessibility needs.