A new method for searching patterns in data using interval estimates is proposed. The computational complexity, algorithmic implementation, and some properties that allow the method to be applied to high-dimensional data are described, including the characteristics of ordering the studied indicators in groups obtained from the analysis, the unambiguous definition of such groups, and their lack of intersections. The possibility of using a generalization of Borda's rule for interval estimates in pairwise comparison of interval estimates in the practical implementation of this method has been studied. The term "ordinal-interval pattern clustering" is proposed. The potential for reducing computational complexity by introducing additional constraints is presented. Recommendations on the appropriateness of using ordinal-interval pattern clustering are discussed. An example of a practical application was studied using synthetic data. Recommendations for using ordinal-interval pattern clustering will help researchers and analysts optimize the data analysis process and obtain more accurate and interpretable results. Examples of practical application of the method on various types of data highlight its universality and effectiveness.