Passenger demand forecasting in railway transportation

Transportation, which has been an important activity throughout history, is very important both socially and commercially. Transportation is important not only in the commercial sense but also for people to move from one place to another. For this reason, passenger transportation is an activity that facilitates social life. In Turkey, with the rapid development of railways, particularly with the development of technology, high-speed train transportation is increasing. Passenger demands are changing due to the fact that rail transportation is a more economical mode of transportation compared to road transportation and with the development of high-speed trains. In this regard, demand forecasts related to railway passenger transportation have gained importance in recent years. In the study, transportation, one of the main logistics activities, is emphasized, the importance of railways and the high-speed train, which has been developing rapidly in recent years, and then passenger demand forecasting was made. Although studies on passenger demand forecasting can be found in both domestic and international literature, no studies spesific to high-speed trains have been encountered in the domestic literature. In the application section of the study, passenger demand forecasting was conducted by using high-speed train data (price, travel time, number of past passengers, numarator, holiday, weekend) between two cities (Ankara–Istanbul) for the period 2021–2024 obtained from the Turkish State Railways (TCDD). In 2024, data requested from TCDD was analyzed using time series Holt's Method, Winters' Multiplicative, Winters' Additive, Simple Seasonal, Multiple Regression and Box Jenkins Analysis (ARIMA). According to the results of the analyses, a relationship exists between the number of passengers and ticket prices as well as the number of services. An increase in the exchange rate leads to higher ticket prices. In addition, it has been concluded that weekends, holidays, the number of services, and time collectively increase the number of passengers. According to the MAPE criterion, the Holt–Winters Multiplicative Model provides the best forecasting performance, whereas according to the RMSE criterion, the ARIMA model provides the best results.