DEVELOPMENT OF A CONCEPTUAL MODEL OF OPERATIONAL – ANALYTICAL DATA MARTS

UDC 004.67
DOI:10.26102/2310-6018/2019.27.4.002

A.P. Raevich, B.S. Dobronets


The storages integrated with analytical systems are focused on dimensional data modeling, which provides quick execution of analytical queries, but has significant drawbacks when working with big data. The article proposes an approach to constructing a conceptual model of operational-analytical data marts, which allows combining the concepts of operational data marts and analytical data marts. Operational data marts are information slices of narrowly focused, thematic information, which designed to solve the problem of operational access to big data sources through the consolidation and ranking of information resources in terms of demand. In contrast to operational data marts that are dependent from sources, analytical data marts are considered as independent data sources created by users in order to provide data structuring for the tasks being solved. The paper provides a comparison of approaches to the construction of analytical queries based on linear queries and associative relationships. The results obtained in this work are used in building a BI cluster on the basis of fast design, analytics, development and implementation of business process models which are performed with using prepared operational-analytical data marts.

Keywords: business intelligence systems, operational-analytical data marts, associative data model.

Full text:
RaevichDobronets_4_19_1.pdf