Mladen Babić (2008) DATA WAREHOUSE FOR BANK TRANSACTION ACCOUNTS BUSINESS ANALYSIS. EngD thesis.
In this work we present implementation of a data warehouse prototype for bank transaction accounts business analysis. As an introduction to extensive data warehousing area we provide some theoretical bases, which include basic terminology, a description of essential elements of data warehouse as well as a description of development methodology we use. We proceed with designing dimensional models of the data warehouse. We design three dimensional data models, which represent bank transaction accounts business activities. We also show integration of designed models into complete bank product analysis solution. When designing models, we use literature recommendations for designing heterogeneous products star schemas, transaction star schemas and periodic snapshot star schemas. By using these recommendations, we design dimensional models, which enable transaction accounts business analysis on different grain levels. The next step in the process of making the data mart is carrying out the ETL system, part of the system that is responsible for refreshing the data warehouse. We develop ETL processes, module for automatic execution of the ETL processes, and module for ETL process error handling. At the end of our work we provide some use cases of the data warehouse for bank transaction accounts. Possible upgrades and guidelines for the future developments of the data warehouse for bank transaction accounts are given in the conclusion.
Actions (login required)