In relation to big data analysis, visualization of multivariate data is helpful in detecting mutual dependencies, characteristic patterns, clusters, and outliers among data and plays an important role in data analysis in various fields is. The Chernoff face, which is one of the multivariate data visualization methods, is a very useful method that can express each observation data with different faces by associating each variable of multivariate data with facial features such as eyes, nose, and mouth of human face, There is no standardized standard for determining which variable is most suitable to correspond to which part of the face. In this study, we propose the use of correlation coefficient to obtain the statistical validity of Chernoff face. In other words, it was found that the correspondence of the variables with high correlation coefficients to the faces of the face showed that the face of Chernoff was expressed to the most purpose. As a result, the following was found. First, we can confirm the validity of existing results to some extent. Second, when visualizing the multivariate data of the big data side through the Chernoff face, it was found appropriate to correlate high correlation variables with the mouth, eyes and nose of various parts of the face.