We are studying the thermodynamic behaviors of social systems . For finding thermodynamic variables which would exist in social systems , we analyzed Japanese regional societies by the principal component analysis . The methods and results of the analysis are presented. The data used are cross-sectional and of time series from 558 cities. We could extract 5 factors through cross -sectional analysis and that of time series. These factors are considered to be important to explain the behaviors of the society. The properties of the factors are discussed from several viewpoints.