They are often images and are used to capture data groupings of importance and more importantly how they relate to one another. Logical Model Logical data models represent the abstract structure of a domain of information. 7Ĩ Artifacts - Data Architecture Continued This model shows the highest level of data groupings giving the users the “concept” of the data structure without extraneous details. It is normally intended for executives or other business users that do not need the complex details of the logical and physical model. Marts CDW ODS 6 6Ĭonceptual Model The conceptual model is the highest view of the data model that is normally a picture. This analysis is normally performed after implementation to resolve data quality or load issues or for future enhancements. Using the example about, the Bottom Up view would be Datamart <- CDW <- ODS <- Source. This is where you start at the final destination point of the data element and trace it backwards to its initial source. ODS CDW Marts 5 5Īnalysis Bottom Up The inverse of Top Down analysis is Bottom Up or Data Lineage. This analysis is primarily done at the creation of the data model to ensure straight thru processing and eliminating duplication of the same data (aka single point of truth). This flow of Source -> ODS -> CDW -> Data Mart is a Top Down view. This data will flow into a Corporate Data Warehouse with the nightly batch cycle and into Data Marts.
This data is sent to an Operational Data Store (ODS) in real time. For example, a piece of data could begin at an user making a transaction on a website. Analysis of this flow is called “Top-Down” analysis. 4 4Īnalysis Top Down Data is perceived to flow downhill or from its initial source or creation point to the final data store destination. Requirements of these architectures are carefully analyzed so that the data store can be optimized to serve the users while achieving a “single point of truth” for each business data element. The physical architecture is the actual database design and data definition language (DDL) that is used to define and create the data structures in the database. Once the logical configuration is defined, the physical Data Architecture is designed to implement it. The logical architecture is a configuration map (picture) of the necessary source data stores including but not limited to a central Enterprise Data Store, an optional Operational Data Store, one or more individual business area Data Marts, and one or more Metadata stores. The Data Architecture is the logical and physical foundation on which the data store will be built. Data Architecture The Data Architecture organizes the sources and stores of business information and defines the quality and management standards for data and metadata. Determine which business subject areas provide the most needed information. Define and prioritize the business requirements and the subsequent data needs the data store will address Identify the business directions & objectives that may influence the data & application architectures. Requirements for a Data Model Establish the scope of the data store and its intended use. All of these items start with quality requirements and a solid data model. The applications provide the business community with access to integrated and consolidated information from internal and external sources. The data management and business intelligence (BI) applications are designed to support executives, senior managers, and business analysts for regular reporting to making complex business decisions. The data management environment positions a business to utilize enterprise-wide operational and historical data to link information from various sources and make the data accessible for a variety of corporate wide user purposes. 1 Our Expertise and Commitment – Driving your Successĭata Modeling Overview DecemOffices in Boston, New York and Northern VA 1Ģ Table of Contents Data Modeling Overview Artifacts Data Architectureĭata Lineage Data Dictionary Tools Modeling Extract, Transform and Load Data Governance Data Stewardship 2 2Ī proper data management programs is a corporate wide strategy and process for building decision support systems and a knowledge-based applications architecture and environment that supports everyday decision making and long-term business strategy.