Data from induction loops, local weather, historical ticketing and calendar data, were used to build operational traffic intensity models to predict traffic at toll plazas. Sharing forecasts with drivers allows them to make departure time changes and avoid congestion.