The Data Science Group in School of Computational Sciences at KIAS investigates the large-scale data-sets of macroscopic complex systems in various disciplines to derive their underlying mathematical and physical principles. Macroscopic complex systems consist of a large number of interacting elements for which their full Hamiltonian or dynamical equations at the microscopic level are either hard to write down or not so useful for deciphering their emergent behaviors, contrasted to the systems of high degree of symmetry or homogeneity. The examples are biological cells, ecological communities, social networks, and economic systems. To extract quantitative and verifiable principles of those systems from the empirical data analysis and iteratively guide the latter, it is essential to develop mathematical models, for which the fundamental theoretical results and methods of statistical physics and network science are applied, such as the mean-field theory, stochastic processes, critical phenomena. and non-equilibrium statistical mechanics, Currently, the research topics of the group are wide ranging, including the intricate wiring of chemical reactions in the cellular metabolism, the interspecific interactions in ecological communities, human mobility, the spatiotemporal patterns of disease spreading, the value distributions in international trade, and so on.
Figure 1. a. Empirical portions of distinct product categories in world trade over years. b. A plant-pollinator mutualistic network model evolving under a non-zero exploitative competition. c. Simulated evolution and speciation of metabolic networks. Each species either expands its metabolic network or gives birth to a new species. d. Empirical plot of the number of metabolic reactions in a species and the species’ evolutionary age.