Profile driven transformation service at the Dutch Central Bank
summary
Centralized transformation service converts the bank's internal DB XML into five outgoing syntaxes (XML, XBRL, SDMX, JSON, CSV) for about 400 reports, using a small domain-specific XML template language plus mapping tables and 5 XSLT stylesheets that run as streaming transforms so they handle streaming large XML up to 1–2 GB. The system enforces semantic-only templates (business defines mappings and groupings), keeps syntax handling in IT, is live in production, and reduced many bespoke converters — training and XSLT support remain operational issues.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:37Decentralised organisation
- 3:00Business reporting
- 5:40One central format
- 6:17Separation of concerns
- 7:22IT department
- 10:53No rocket science
- 15:03Mapping
- 15:30SDMX
- 16:30Semantic naming
- 17:09Grouping
- 18:32Multipart grouping
- 19:11Constant pressure
- 19:42Streaming accessibility
- 21:23XSL fork
- 22:48Transformation service
- 25:31Observations
- 30:07Future plans
related talks
29mManaging your Data Assets: The interview
20mModernizing data architecture with data products in asset management
31mMetadata-based Dynamic ETLs
30mData Mesh – Architecture And Insights From The Self-service Platform Boosting Financial Services
39mScaling ETL Migrations to 300x with AI
24mRetail Transactions at scale for AI - From PoC to production
49mBuilding a Data-Driven Bank
27mDemocratising data-driven decisions with self-service tools