DataWorks:Data modeling overview

Prerequisites

Intelligent Data Modeling is a value-added service in DataWorks. Activate it before use. For more information about different editions and billing details: Intelligent Data Modeling Billing.

Limitations

Overview

DataWorks Data Modeling covers Data Warehouse Planning, Data Standards, Dimensional Modeling, and Data Metrics. Materialize Dimension Tables, Fact Tables, and Aggregate Tables from your designs into a compute engine.

• Data Warehouse Planning

  Design Data Layers, Business Categories, Data Domains, Subject Areas, and Business Processes on the Data Warehouse Planning page.

  • Data Layer

    Design Data Layers based on your business and data scenarios. DataWorks provides a default five-layer architecture:

    • Operational Data Store (ODS)

    • Data Warehouse Detail (DWD)

    • Data Warehouse Summary (DWS)

    • Application Data Service (ADS)

    • Dimension (DIM)

    You can also create custom Data Layers. Create a Custom Data Layer.

  • Business Category

    Define Business Categories to help different business types quickly locate their data, then associate categories with Dimension Tables and Fact Tables during modeling. Business Category.

  • Data Domain

    A Data Domain is a high-level data classification standard—a collection of abstracted business processes that helps users quickly locate relevant data.

    Each Data Domain corresponds to a macro-level analytical area such as procurement, supply chain, HR, or e-commerce. A unified team or individual, like a data architect or modeling team member, with deep business understanding should manage Data Domains. Data Domain.

  • Business Process

    A Business Process describes a business activity flow. In e-commerce, adding to cart, placing an order, and payment are each a Business Process. A common use case is funnel analysis—tracking orders at each stage (browsing, cart, order, payment, receipt) to analyze the "order count" metric throughout the funnel. Business Process.

  • Data Mart

    A Data Mart defines detailed business themes for a specific Business Category, using Subject Areas to partition data by analytical perspective to serve business applications and statistical analysis. An example is a Data Mart for an operations platform. Data Mart.

  • Subject Area

    A Subject Area divides a Data Mart by analytical perspective—a collection of closely related data themes grouped by business focus. Common examples in e-commerce: transactions, members, and products. Subject Area.

• Data Standard

  Define Data Standards before modeling or develop them from business practices over time. Standardize Lookup Tables, Measurement Units, Field Standards, and Naming Dictionaries to ensure consistency across modeling and application.

  For example, a registration table stores member ID in a field named user_id, while a logon table uses userid. Create a unified Field Standard for member ID that specifies a Lookup Table, field properties (data type, length, default value), and Measurement Unit. Associate this standard with the member ID field in all models to ensure cross-table consistency.

  Create Field Standards in DataWorks: Field Standard.

• Dimensional Modeling

  DataWorks follows Dimensional Modeling principles. When designing a data warehouse with Dimensional Modeling:

  • Dimension Table

    Extract potential dimensions from your Data Domain plan and store them and their attributes in Dimension Tables. For example, in e-commerce: the order dimension (order ID, creation time, buyer ID, seller ID), user dimension (gender, birth date), and product dimension (product ID, name, launch time). Deploy these tables to your data warehouse and use ETL jobs to populate them with actual dimension data.

  • Fact Table

    Identify factual data generated during each Business Process and store it in Fact Tables. For example, for "place order," create a Fact Table recording order ID, creation time, product ID, quantity, and amount. Deploy these tables and use ETL jobs to consolidate real data for business analysis.

  • Aggregate Table

    Pre-aggregate detailed fact and dimension data into an Aggregate Table based on analysis needs and data warehouse layers. Query the Aggregate Table directly instead of the more granular Fact and Dimension Tables.

  • Application Table

    An Application Table organizes multiple Atomic Metrics and Derived Metrics that share the same Period and dimensions for a specific business scenario. It serves as a foundation for business queries, OLAP analysis, and data distribution.

  • Reverse Modeling

    Import models generated by other tools into the Dimensional Modeling module. If you have existing models and want to switch to DataWorks intelligent modeling, Reverse Modeling quickly imports them without remodeling.

  Create logical models: Create a logical model: Dimension Table, Create a logical model: Fact Table, Create a logical model: Aggregate Table, and Create a logical model: Application Table. Reverse modeling: Reverse Modeling: Reverse a Physical Table.

• Data Metrics

  Data Modeling provides a Data Metric feature to build a unified metric system.

  The metric system consists of Atomic Metric, Modifier, Period, and Derived Metric.

  • Atomic Metric: A measure based on a Business Process, such as "payment amount" in the "make payment" process.

  • Modifier: A qualifier that limits the scope of a metric, such as limiting the "payment amount" statistic to "maternal and infant products".

  • Period: A time range or point for a metric, such as specifying the "payment amount" to be calculated over the "last 7 days".

  • Derived Metric: A combination of an Atomic Metric, a Modifier, and a Period. For example, "total payment amount for maternal and infant products over the last 7 days".

  Build a metric system: Data Metric.

Why data modeling matters

• Standardized Management of Massive Data Volumes

  As businesses grow, data volume and structural complexity increase rapidly. Managing and storing this data in a structured and orderly way is essential.

• Break Down Data Silos and Enable Connectivity

  When business units manage data independently, silos form and prevent decision-makers from getting a clear and quick overview of company-wide data. Breaking down these silos is a key challenge in enterprise data management.

• Unified Data Standards for Flexible Integration

  Inconsistent data descriptions cause management difficulties, content duplication, and inaccurate results. Unified Data Standards enable flexible integration with upstream and downstream services without disrupting existing architecture.

• Maximize Data Value and Business Profit

  Effective Data Modeling maximizes enterprise data value, delivering more efficient data services for your business.