CFD simulations have been utilized to improve a hull form. As a result, a huge amount of numerical data has been acquired. In the verification process of numerical results, comparison among data obtained from different kinds of resources (e.g. experiments, other numerical experiments) plays an important role. Remarking the comparison process among data, it is found that the most difficult issue is the difference of locations that the data is defined. In this paper, we try to solve this problem. First, drawbacks of image data for the comparison in a wide spatial region are clarified. In order to overcome the drawbacks, a concept of the comparison space is introduced. Second, how to construct the comparison space is described. Finally, several numerical experiments are demonstrated to validate our concept. In the case of the linear advection problem, a new characteristic feature of the numerical error is found. In additions, it is shown that the grid dependency on the numerical results can be found out in the wide spatial region.