Resumen
This article is devoted to the analysis of traffic flows based on correspondence matrices. Such matrices describe the number of displacements between two points for a certain time interval. From a practical point of view, data relating to the Moscow Metro are considered. Accordingly, the correspondence matrix describes the movement between stations. Theoretically, such data describes all the characteristics of passenger traffic. In practice, it depends, of course, on the selected data processing model. Often, such matrices are used only for simple statistics, such as the number of passengers transported over time. At the same time, the space-time information interesting for digital urbanism is lost. For example, how were the trips distributed over time, how stable are these distributions, etc. The paper provides a detailed review of existing approaches to the analysis of data in correspondence matrices. As a practical task, a short-term forecast of passenger traffic is considered. It is noted that the short-term traffic forecast is a challenge that has been the subject of many research papers in the past few decades. Most of the work has been historically devoted to the analysis of traffic flows exclusively by road transport. The study of railways and, in particular, underground transport with its specificity has long been ignored. Relevant studies have been conducted only recently.