As an approach to the analysis of complex systems such as social, economic and environmental systems, a lot of complex and ambiguous multidimensional data must be treated. Therefore, a technique of data presentation suited to the visual sense is required. But there has been almost no systematic study of expressing data, because even in systems engineering, not much attention has been paid to this area of study.
This paper describes the use of a face pattern, which is illustrated by a newly developed method, for expressing the various states of a complex system and their processing. This face pattern is based on the results of the analysis of facial expressions. Some technique of assigning data components to facial features, which make use of the characteristics of facial expressions, are proposed in this paper. As a practical application of the present Faces Method, the emission data of six air pollutants of Tokyo's 23 wards are expressed in the form of various facial expressions, which are classified by a developed classification procedure, and the resulting characteristics of the emission data are extracted from the facial expressions. These results are compared with the results obtained from cluster analysis and principal component analysis. Moreover, one dimensional scaling of the classified facial expressions is tried to evaluate the emission data. The efficiency of the Faces Method is investigated through these experiments.