Masterclass Data Warehouse Design & Data Management.195 Add to favorites

A data warehouse is a data source for users in Business Intelligence environments. A good data architecture is essential for quickly delivering data reporting or extraction.
We also want to minimize disruptions to the data warehouse from updates, maintenance, or other users. A layered model is useful in this case, as are temporary storage areas like the Operational Data Store. Technically, we can also specify layers to separate near-real-time data from batch-oriented data, increasing processing speed.

Large amounts of data and many layers require a dedicated process to extract data (E), transform it (T) into the correct data models in the various layers, and transport or load it (L). These so-called ETL processes must be well-defined, with agreements on how and when they are executed. ETL execution also serves a quality assurance function; we want the correct data in the data warehouse, not vague, incomplete, outdated, or inconsistent data. We'll examine data cleansing, profiling, and enrichment techniques.

Finally, the nature of the data warehouse: do we use vector techniques like EXCEL and a multidimensional data warehouse (MOLAP) or the familiar read/write techniques used in relational databases (ROLAP). Or perhaps both? And how do we start? Do we operate event-driven/data-driven, or do we ask the future BI user (requirements-driven). Or do we base our approach on management objectives (goal-driven), or perhaps a combination of these approaches?

In this 5-day masterclass, lasting 11 to 15 weeks, we'll explore the concepts and techniques mentioned above and examine which models and requirements can be used to set up a set of data marts from scratch in a project. We'll also look at how to maintain and use these for BI reporting and the information needs of various end users.

Topics
- Roles in BI & Data Warehousing
- Data Governance
- Data Mesh
- Meta data, master data and reference data
- Business rules identification.
- Business event driven approach
- Requirements-driven approach
- Data-driven approach
- Goal-driven approach
- Conceptual modeling with ERD-ISO/ISD.
- Logical modeling with the relational model.
- Data Fact Modeling
- Attribute trees
- NIAM-FCOM/Object Role modeling
- Convert business rules into information rules for the data model.
- Information rules implementation in ERD.
- Logical/conceptual modeling of data warehouses
- Conceptually modeling a data warehouse with Data Fact Modeling.
- Conceptual Modeling with Data Vault.
Logical Data Vault design
From Data Vault to Dimensional Models
- Business Process information needs assessment
- Workload & Materialized views
- Star models and Snowflake models.
- Data Mart design
- Anchor modeling
- From Achnor models to Dimensional models
- ETL processes
- Data profiling
- Data cleansing
- Data enrichment
- Operational data store
- Data lake
- Data lakehouse

Learning objectives

• The participant can perform factual analysis, identify relationships and rules between entities at various levels of detail, including conceptual and logical models.
• The participant understands data modeling via Object Role Modeling (FCOM) and can express fact types, understand uniqueness rules and constraint rules and the transformation to a relational model.
• The participant knows how to create an information structure diagram (ISD) or ERD.
• The participant knows how to use top-down and bottom-up approaches to initiate development of a data model and then work out the middle-out.
• The participant knows how to specify metadata.
• The participant can identify master data and reference data.
• The participant can apply data normalization in an ER model and relational database schema.
• The participant knows various techniques to discover additional requirements and implement business rules in the data model.
• The participant can 'read' a data model and convey it to subject matter experts, using the associated terminology.
• The participant understands how Data Fact modeling and Data Vault modeling works.
• The participant has knowledge of multi-dimensional databases and hyper(cubes).
• The participant understands how to determine a workload from reporting needs for a data warehouse.
• The participant understands how to define queries for reporting that fit within a multi-dimensional database environment.

Investment training
See cost specification. There are additional material and lunch costs, as well as any exam costs.

Walkthrough
11 to 15 weeks, one lesson day every 2 to 3 weeks in principle.

Level/Prior knowledge
A higher professional education (HBO) level of knowledge and some knowledge of relational models and SQL is required. This is a postgraduate course.

Exam
The Certified Professional Data Warehouses* exam consists of 30 multiple-choice questions, takes a maximum of 3 hours, and is an open-book exam.

* the Certified Professional Data Warehouses assessment is fully owned and maintained by The Business & IT Exams Association.