Forecasting of public bus passenger flow in the Voronezh region

Introduction. The relevance of forecasting the passenger flow of public bus transport in the conditions of the steady decline in the volume of traffic, the aging of rolling stock and the transformation of population transport priorities has been substantiated. The need to develop a transparent, interpretable and statistically valid model, considering both demographic and infrastructure factors, has been shown. The purpose of the study is to develop an interpretable and statistically substantiated regression model of medium-term prediction of the passenger flow of public bus transport, taking into account demographic and infrastructure determinants, with the adaptation possibility to the conditions of different regions. Methods and materials. The study uses official statistical data of the Voronezh Region for the period from 2010 to 2024, including the number of transported passengers, the population and the availability of buses in operation. Multiple linear regression based on a time trend is used as a forecasting method. Methodology. A regression model of the dependence of passenger flow on demographic and infrastructural factors has been built; the parameters have been evaluated using the least squares method; the quality of the model has been checked by the coefficient of determination, the statistical significance of the coefficients and the analysis of residuals. Results. The model has demonstrated a high explanatory power (R² = 0.94); the forecast for the period from 2025 to 2031 indicates a steady decrease in passenger traffic – from 254.7 thousand to 177.5 thousand people, which is primarily due to the demographic decline. Discussion. The negative coefficient for the number of buses does not reflect a cause-and-effect relationship, but rather the system’s response to falling demand; the positive trend compensates nonlinear effects of recent years. The results coincide with national trends and emphasize the need to move from quantitative to qualitative transportation management. Conclusion. The developed model is interpretable, practically applicable, and suitable for strategic planning. In the future, it is planned to expand the set of variables, switch to route-oriented forecasting and integrate the model into the system of scenario planning.