Data Preparation for Advanced Data Analysis on Elastic Stack

This paper presents approaches for preparing different types of data to be loaded into the document-oriented NoSQL Elasticsearch database. The considered database allows not only to store data, but also provides an opportunity to use Kibana, data visualization utility, which is a powerful tool for data analysis. The task of preprocessing is essential, because well-prepared data not only allows you to increase the accuracy of the analysis, but also expand its capabilities. For more coverage, the approaches are described with the use of real cases that have been solved by analysts. The paper presents methodological and practical ways to solve problems both by transforming the data and adding new fields, and by correctly mapping for Elasticsearch indexes. For a clear demonstration of the approaches, their practical application is given on the example of two datasets with bibliographic information on papers and information on funding of scientific and technical projects. The demonstration shows the difference between initial and enriched data, as well as the charts built by working with the data, which enables advanced data analysis.