Advances in intelligent transportation systems have resulted in deployment of surveillance systems that automatically collect and store extensive networkwide traffic data. Dynamic traffic assignment (DTA) models have been developed for a variety of dynamic traffic management applications. They are designed to estimate and predict the evolution of congestion with detailed models and algorithms that capture travel demand and network supply and their complex interactions. The availability of rich time-varying traffic data spanning multiple days provides the opportunity to calibrate a DTA model's inputs and parameters offline so that its outputs reflect field conditions in future offline and online real-time applications. The state of the art of DTA model calibration is a sequential approach, with supply model calibration (assuming known demand inputs) followed by demand calibration with fixed supply parameters. An offline DTA model calibration methodology is presented for simultaneous estimation of all demand-and-supply inputs and parameters, with sensor data. A minimization formulation that can use any general traffic data and present scalable solution approaches for the complex, nonlinear, stochastic optimization problem is adopted. A case study with DynaMIT, a DTA model with traffic estimation and prediction capabilities, is used to demonstrate and validate the methodology. Archived sensor data and a network from Los Angeles, California, are used to demonstrate scalability. Results indicate that the simultaneous approach significantly outperforms the sequential state of the art in terms of modeling accuracy and computational efficiency.