Abstract

Users of a business intelligence (BI) system employ an approach referred to as online analytical processing (OLAP) to view multidimensional data from different perspectives. Query languages, e.g., SQL or MDX, allow for flexible querying of multidimensional data but query formulation is often time-consuming and cognitively challenging for many users. Alternatives to using a query language, e.g., graphical OLAP clients, parameterized reports, or dashboards, are often not a full-blown alternative to using a query language. Experience in cooperative research projects with industry led to the following observations regarding the use of OLAP queries in practice. First, within the same organization, similar OLAP queries are repeatedly composed from scratch in order to satisfy similar information needs. Second, across different organizations and even domains, OLAP queries with similar structures are repeatedly composed from scratch. Finally, vague requirements regarding frequently composed OLAP queries in the early stages of a project potentially lead to rushed development in later stages, which can be alleviated by following best practices for OLAP query composition. In engineering, knowledge about best-practice solutions to frequently arising challenges is often documented and represented using patterns. In that spirit, an OLAP pattern describes a generic solution for composing a query that allows a BI user to satisfy a certain type of information need given fragments of a conceptual model. This paper introduces a formal definition of OLAP patterns as well as an expressive, flexible, and generally applicable definition language.

Highlights

  • Multidimensional data analysis is central to descriptive analytics in general and online analytical processing (OLAP) in particular

  • In this chapter we present a proof-of-concept prototype implementation of the pattern-based approach to multidimensional data analysis that is based on the definitions and grammar for OLAP patterns defined in the previous chapters, following an architecture that we introduce in this chapter

  • In the agriProKnow project, OLAP patterns were originally introduced to be able to deal with uncertainties regarding the stakeholders’ requirements with respect to OLAP queries to be answered by the data warehouse

Read more

Summary

Introduction

Multidimensional data analysis is central to descriptive analytics in general and online analytical processing (OLAP) in particular. We present a pattern-based approach to multidimensional data analysis, the development of which was informed by experience from cooperative research projects with industry participation. We describe the motivation behind the pattern-based approach to multidimensional data analysis (Section 1.1), sketch the presented approach (Section 1.2), explain the contributions of this thesis (Section 1.2), and give an overview of the organization of the remainder of this thesis (Section 1.4)

Objectives
Results
Discussion
Conclusion
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call