After specifying business processes in each data domain, you need to define dimensions and build a bus matrix based on these dimensions.

Define dimensions

When dividing data domains and building a bus matrix, you need to define dimensions based on the analysis of business processes. This tutorial takes the e-commerce marketing business unit of Company A as an example. In the transaction data domain, this tutorial focuses on the business process of receipt confirmation, where a transaction is successful.

In the business process of receipt confirmation, dimensions can be defined from two perspectives: product and place of receipt. This tutorial assumes that each product is purchased and received in the same place. From the product perspective, the following dimensions can be defined:
  • Product ID, which is the primary key
  • Product name
  • Product price
  • Product status: 0: new; 1: idle; 2: second-hand
  • Product category ID
  • Product category name
  • Commodity ID
  • Commodity name
  • Buyer ID
  • Product sales status: 0: normal; 1: deleted; 2: taken off the shelf; 3: never put on the shelf
  • City of the product
  • Province of the product
From the place of receipt perspective, the following dimensions can be defined:
  • City code
  • City name
  • Province code
  • Province name

As the core of dimensional modeling, dimensions must be unique in an enterprise-level data warehouse. For example, each dimension defined from the product perspective must be unique for Company A. The province code dimension delivers the same information in any business process.

Build a bus matrix

After specifying business processes in each data domain, you can build a bus matrix. You need to identify the dimensions that are related to each business process and define the business processes and dimensions in each data domain. The following table lists the bus matrix of the transaction order feature in the e-commerce marketing business unit of Company A. The dimensions defined include the province, city, category name, category ID, brand name, brand ID, product name, product ID, and turnover.
Data domain/Business process Conformed dimensions
Province City Category ID Category name Brand ID Brand name Product ID Product name Turnover
Transaction Order placement Yes Yes Yes Yes Yes Yes Yes Yes No
Payment Yes Yes Yes Yes Yes Yes Yes Yes No
Shipping Yes Yes Yes Yes Yes Yes Yes Yes No
Receipt confirmation Yes Yes Yes Yes Yes Yes Yes Yes Yes
Note Yes indicates that the dimension is included, whereas No indicates that the dimension is excluded.