In this study, we present a numerical method to analyze the dynamics of the Russia-Ukraine war at the local level using data-driven system identification. The target dataset comprises losses in armored equipment from both forces, weighted based on the types of equipment. We begin by first upsampled the data using a monotonicity-preserving subdivision scheme. Next, we modeled a linear system of ordinary differential equations for each local region from the upsampled data through least-squares fitting. By investigating the behavior of coefficients of these local systems, we analyzed the warfare dynamics. We present simulations of the proposed method to validate its effectiveness.