In this paper, firstly, we propose to use time series analysis to observe the development and change process of social events, using the DF test and ADF test to eliminate the random autocorrelation in order to exclude the interference of white noise series. Secondly, the instantaneous frequency of non-smooth data is defined by analyzing the signal, and the constraints are modified from global constraints to local constraints. The Fourier-transformed eigenmode function is used to satisfy the four necessary conditions of completeness, orthogonality, localization, and adaptability. Finally, the development trend of ethnic folk dances and the change in communication dynamics are analyzed. The results show that the change of language on the spreading attitude of folk dance is 4.381, the change of custom on the spreading attitude is 5.943, the change of culture on the spreading attitude is 10.866, and the change of culture on the spreading attitude of folk dance is the largest, which shows that the development of folk dance away from the development of folk dance is not separated from the support of culture. The study will have a positive impact on passing on the distinctive art of ethnic folk dance and enriching Chinese traditional culture and art.