For efficient air traffic management, it is essential to understand the traffic characteristics of the current operation and to be able to predict the traffic volume and capacity. In this paper, the change in aircraft count in all sectors and terminal maneuvering areas in the Incheon Flight Information Region is studied using one-year trajectory data. After obtaining the distribution of traffic volume change in one day, the maximum allowed aircraft was obtained for all airspaces. For further investigation, a machine learning-based clustering technique, Gaussian Mixture Model was used to cluster the change in aircraft count for each airspace. The results show that for each airspace, the daily change in aircraft count follows several patterns. Finally, it was shown that a certain level of prediction can be made to predict traffic changes and capacity from given traffic data by comparing it with the clusters.