Data Mining Efficiency in the ESG Indexes Verbalization Analysis (on the Example of the MSCI Site)

The authors consider the efficiency of the data mining method for analyzing the verbalization specifics of the ESG indexes on the example of business documentation and correspondence of MSCI Inc. as one of the best-known global suppliers of ESG and climate products. The relevance of the research is due to its interdisciplinarity and testing of new methods for large text corpora analysis. Text mining implies methods for obtaining new information from the flow of texts or large text data and involves compilation, organization, and analysis of large collections of documents to extract the necessary information and detect previously unknown links between texts. A three-step automatic selection to determine the most widespread lexical patterns when analyzing the verbalization of the ESG indexes demonstrates that these steps, namely text segmentation, n-gram detection, and category labeling, are related to each other. The scientific novelty of this research is determined by the flexibility of the proposed analysis model because it can be used as an open-source tool, i.e., the approach presented herein can be extended to many other types of texts and languages.