I. Introduction
An incredible amount of data is created every day within transportation. Technology such as modern-day smartphones, for example, are carried by millions of people, and are often equipped to capture movement patterns through accelerometers, gyroscopes, or GPS traces. There are cameras that overlook street intersections, loop detectors embedded along roads, train stations that record entry and exit traffic, and logs of individuals requesting rides to and from specific locations in taxi or ride-sharing scenarios. Many modern transportation systems have begun to embrace the idea of data-driven paradigms as a means to obtain more accurate predictions or advanced control policies, by learning from this data [116].