Abstract: This paper proposes an improved forecasting approach for predicting subway ridership in New York City by integrating calendar effects such as holidays and weekends into Seasonal Exponential ...