[Abstract] This article shares the practical experience of bank data warehouse construction, including construction ideas, hierarchical planning, model design, subject division, auxiliary tools and practical thinking, hoping to provide reference for peers who are engaged in similar project construction.
In recent years, with the increasing development of banking business and the deepening of regulatory requirements, Bank of Qinhuangdao has gradually taken data analysis, data application, and data mining as important supporting means for financial business development and management decision-making, taking data as the core asset of the bank, and continuously expanding data Application scenarios to improve data asset management capabilities. The chimney system construction model will bring serious consequences to data construction. Data standards and data norms cannot be implemented as expected, and the improvement of data quality is far from expectations. Qinhuangdao Bank's original data platform (ODS for short) system can no longer meet the increasingly vigorous data analysis needs of the business department, and is also facing some other problems, such as: chaotic data structure, scattered data storage, serious data redundancy, and data network structure, lack of a unified data model and data management and control, etc., there is an urgent need to restructure the ODS system, integrate data in a unified manner, and build a bank-wide data standardization system to meet the needs of rapid business development and data-assisted applications in the bank. Play its due value in the competition of digital transformation.
1. Construction ideas
Qinhuangdao Bank officially launched the data warehouse upgrade project construction in July 2021. The project team proposed a data warehouse construction method of "data standards first, data control follow-up, application-driven and data-driven combination" during project construction. Data standards, Data control and data application are unified into the scope of data warehouse construction. As the basic data base of the whole bank, the data warehouse is connected to 44 business systems, and 13 business theme models are designed according to the DW five-tier architecture with business-driven planning, and the unified basic data resource management is realized in accordance with the principle of "one number, one source" to avoid duplication Construction and index redundancy ensure the standardization and unification of data standards, realize the full link association of data assets, and provide data support for data analysis.
-
Data Standards First
The data standard is a