How many times have you waited for an H bus only to see three LX’s roll by? And when the H finally arrives, it’s so packed that you can’t get on? Have you ever wondered how the university decides how many buses to run, and when? The answer is by building models of student bus travel. Models, and algorithms based on them, actually permeate our lives, from weather forecasts to political predictions to online shopping and streaming. In this seminar, we will examine models and algorithms to see how they work, how they are built, and how they can be “scientific” even if they aren’t just about science. Through a series of case studies, we will examine the simplifications and assumptions that are built into models, discuss circumstances in which they are reasonable, and consider how and why they break down. We will then think about our own model of Rutgers bus travel, focusing on the information and assumptions needed to make it work. Ultimately, our goal is to understand that “all models are wrong but some are useful” (George Box) and “algorithms are opinions embedded in code” (Cathy O’Neil).
Charles Keeton (Physics and Astronomy; Academic Dean, School of Arts & Sciences Honors Program)
01:090:101 section 35