Demand for vehicle and public transportation systems continues to increase in and around major urban centers. This increase is especially pronounced during the morning and evening commutes and is further complicated by the complex spatial interactions that influence the variation in system demand. In an effort to help agencies better understand this variability and develop better demand forecasts this research investigated the underlying factors impacting public transportation ridership regardless of transit mode, then uses this insight to estimate specific models to help forecast changes in subway ridership. The spatial database for the case study consisted of social, economic, and land use characteristics for the 2166 census tracts in the five boroughs of New York City, NY. The spatial models were found to have a better overall model fit compared to their non-spatial counterparts. Moreover, spatial dependence was found to be statistically significant in both models.