This thesis is motivated by a practical problem in emergency department (ED) operations management. Prolonged waiting times and overcrowding are prevalent in EDs as a result of the mismatch between demand (i.e., patient arrivals) and supply of ED services (Morley, Unwin, Peterson, Stankovich, & Kinsman, 2018). As the gateway to modern healthcare systems, EDs are faced with the arrival of patients with urgent and complex care needs, increased arrivals of the elderly, and high volume of low-acuity patient arrivals. To improve the operational efficiency and healthcare delivery, ED administrators have to make informed decisions about efficient allocation of resources; demand forecasting is a first step towards informing such decisions. Using a Montreal hospital ED as a basis for our investigation, we evaluate the effectiveness of the rarely used regression with autoregressive integrated moving average errors (regARIMA) model in forecasting future daily and hourly ED arrivals. We also experimentally evaluate the performance of Facebook Prophet (fbprophet) and demonstrate its competitiveness with established forecasting methods. This insight is particularly valuable given that in the ED arrival forecasting literature, it is viewed as a “Blackbox” or “off-the-shelf” method and has not been used for comparison with other established methods. Furthermore, we investigate the hypothesis that public sporting events, particularly hockey, lead to increased arrivals to the ED by using hockey games as a predictor within our forecasting models.