Behind the Course

The people behind this course

A course like this one is built on the work and the generosity of others. I would like to thank five of them.

Professor Kaan Ozbay (NYU Tandon) invited me to teach this course, and his advice shaped how it is built. I first got to know him through a book rather than in person: Feedback Control Theory for Dynamic Traffic Assignment, written with Pushkin Kachroo, which brings the tools of control engineering to the routing of traffic across a network. That reach across disciplines is typical of him, and so is his insistence that models must be grounded in data: a model is only as good as the measurements it is tested against. Both shaped this course, down to its name: data-driven mobility modeling, where classic theory, real data and new methods are told as one story. His impact reaches well beyond one course: he is the founding director of C2SMART, NYU’s federally funded University Transportation Center, and since 2013 a member of NYU’s Center for Urban Science and Progress (CUSP), where research is put to work on the real mobility problems of New York and other cities.

Professor Panos Michalopoulos (University of Minnesota) led me into the world of traffic flow theory, and he has been an inspiration ever since. He is both a great traffic engineer and a traffic flow theorist: he founded Image Sensing Systems, whose Autoscope cameras turned video into traffic data on roads around the world, and he developed KRONOS, a macroscopic traffic simulator that put continuum models of traffic to work on real freeways; that work is still read today. His principle was simple: theory must be applicable to solving real traffic problems. That principle runs through this course, from the first wave on a ring road to the last simulation with a real signal controller in the loop, and his chapter on continuum flow models is among the readings.

Edward B. Lieberman, founder of KLD Associates, was my colleague there, and for years we shared an office. He is one of the greatest traffic engineers I have ever known, and I learned a great deal from him. He is also a recognized pioneer of traffic simulation: from the early 1970s his team built some of the first computer models that follow every vehicle through a city’s streets (UTCS-1 and NETSIM, whose code still runs inside today’s CORSIM). His evacuation model, DYNEV, has been used for the evacuation time estimates of most nuclear power plants in the United States. In 2007 the Transportation Research Board (TRB) honoured him with its first Traffic Simulation Pioneer Award, for lifetime achievement. The second half of this course, where simulation meets real signal controllers, stands on ground he helped lay.

Professor Joseph Chow (NYU Tandon) challenged me to design and build a course that students find interesting rather than intimidating. Much of what makes this one what it is, the puzzles that open each lecture, the demos you can play with, the real data with its surprises, grew out of that challenge. He is Deputy Director of C2SMART and leads BUILT@NYU, the Behavioral Urban Informatics, Logistics and Transport Laboratory, where his research looks at transportation as a whole system: multimodal networks, the mobility services people use and the choices they make, urban logistics, and the economics that tie them together. His work is recognized well beyond NYU, with an NSF CAREER award and editorial roles at leading journals.

Professor Daiheng Ni (University of Massachusetts Amherst) is a personal friend whom I highly respect. He is well known in traffic flow theory for his field theory of traffic, in which each driver moves through a “field” shaped by the road and the vehicles around it, an idea that connects microscopic and macroscopic models in one unified view, and for his Longitudinal Control Model of car-following. His book Traffic Flow Theory (2nd edition, Elsevier, 2026) is the one I used when I last taught this subject and the one this course follows most closely. It does something rare: it takes the reader from what traffic looks like in data, through the fundamental diagram, waves and shocks, to numerical methods and car-following models, as one connected theory, with the derivations written out and nothing waved away. Students who want to go deeper than the lectures will find the next step there, chapter by chapter.

Any errors in these pages are mine.

Wuping Xin