In our previous research, we applied the two-sided Taguchi method to condominiums in an urban area of Tokyo and showed that the effects of green landscape differ depending on the area (riverside area, uptown area, or bay area) in which the condominium is located. A problem with this analysis is that in the development of new real estate, the results cannot be applied to actual develepment cenditions without knowing the type of area in which the real estate will be classified. In this study, we tried to classify each area by using item diagnosis so that the previous research results could be used for practical purposes. With the conventional method, diagnosis took time, as it was hard to judge eiearly whether patterns of areas were alike. By calculating correlation coefficients between items and drawing a scatter chart, we were able to classify the item pattems for each area and obtain appropriate criteria that could be applied to real estate developments of unknown type. In addition, it was shown that this method can provide criteria for classification of anomalies. Although this method is still in the trial stage, it has made item diagnosis more efficient.
