Development of a travel time prediction algorithm is challenging due to its high variability. Urban roads with signalized and unsignalized intersections witness constant formation and dissipation of queues, and platoon formations and dispersals, which make travel time prediction difficult. This study develops and compares data-driven, analytical, and hybrid models to predict the travel time in an urban arterial road section. The seminar will include details of a data-driven model for prediction of travel time at urban arterials and an analytical model for queue length estimation. Data-driven models for travel time prediction in literature have been shown to perform well for highways and freeways. In this study, we investigate the performance of data-driven models at a signalized intersection approach. Since queue length is a major factor that influences the travel time at intersections, methodologies for queue length estimation have been developed to predict travel time. An analytical model was developed to estimate the queue length at signalized intersections from sample travel time data. A hybrid methodology combining a traffic flow theory-based model, namely, the Link Transmission Model, with a filtering technique, namely Particle filter has also been developed for the estimation of queue length. The traffic flow theory component of the model captures the dynamics of traffic, while the data-driven component captures the subtle traffic features that may not be captured by the traffic flow theory component but are necessary for accurate prediction of travel times.