To allow for easily grasping semantic understanding from multidimensional distribution data of event items and attribute items.SOLUTION: A data analysis method comprises: performing information compression of multidimensional distribution data of event items and attribute items by means of machine learning; assigning coordinates to the event items and the attribute items in a same multidimensional space; plotting indicators corresponding to the event items and the attribute items in the same multidimensional space; and clustering the indicators with multidimensional cosine distance so that semantics are presented in a unified scale to perform cluster profiling.SELECTED DRAWING: Figure 3
展开▼