Data RequirementsNeed
to evolve data requirements for an information management requirements
specification and want practical timesaving suggestions?
Task Objectives  The objective of this task is to define requirements for:
Information management, or data warehousing projects, are data centric. This means that most of the requirements will be data related and require highly skilled IT professional specialists early in the project, to define requirements, and thru-out the project, to ensure requirements traceability and satisfaction of requirements. Clearly defined requirements will go a long way towards ensuring project success Structure and design An entity relationship diagram documents the logical data model and metadata e.g. a definition for each entity and attribute. This model represents logical business rules e.g. the model might show a business rule which states “a person can own many automobiles, and an automobile must be owned by one person at one time” The users can look at the model and agree that the logic is correct. The technical team can use the model as the basis to ensure that the data base design will not permit an automobile to be owned by more than one person at one time, or, in this case, not permit storage of an automobile without an owner. A dimensional model documents the facts and dimensions that will be used for analytic and reporting purposes. Think of a fact is something that has happened, e.g. an invoice. Think of a dimension as all of the things that relate to the fact e.g. a product dimension, a customer dimension, a sales territory, and a date dimension. These dimensions are used to analyze details about the fact e.g. it helps answer questions like which products were sold, by which sales territory, to what customers and when. Data storage Data storage requirements are concerned with data volumes and expected growth rates e.g. if we know that we have 100,000 customers today and are expecting a growth of 10% by month, then these data requirements need to be documented to provide direction to the architecture and design team for data base sizing and capacity planning. Data movement Data movement requirements confirm the source data interface information, which will provide guidance to the architecture and design team. If the solution builds a data warehouse, which feeds other systems, then the data movement requirements will also include interface information for the “down-stream” systems. It is very important to get user input on the data movement requirements e.g. sales data might be derived from a sales application or from a billing system. The user team needs to confirm which system is the “system of record” for the sales data i.e. the system that is “trusted” to provide the most accurate data. Data security Data security sounds like a non-functional requirement but needs to involve users. Suppose you have a human resource application that contains sensitive executive salary and bonus information. This application is used by the HR department but will now be loaded into a data warehouse which can be used by a variety of other departments—It is important to identify the security data requirements and consider who can read the data, who can extract it for “downstream” systems, and how to prevent unauthorized access to test data. Data quality Data quality requirements are critical to a successful information management project—it is important to identify how data quality issues will be handled. Things to consider include should we load bad data? Should we reject it? Should we store it until a data steward corrects it? Who should be notified when data quality issues occur? Information usage Business intelligence and analytical reporting requirements should specify which data metric requirements will be used and provide a definition for each metric e.g. if invoice count is defined as a count of all invoices sent, then it is clear that it does not include invoices that are still in progress. This definition helps the architecture and design team build a solution that only extracts invoices with a status of “sent”. Information usage should also consider:
Information management projects are data centric. Most of the requirements are data related and require highly skilled IT professional specialists early in the project, to define requirements, and thru-out the project, to ensure requirements traceability and satisfaction of requirements.
|
|
 |
|
