Data Oriented Analysis

• data are more stable and not changing than the processes that use them.
• data entities and their attributes are rarely changed, event if their values are.
• data administration: conscious management of data as a resource of the business.
• Information Engineering (IE): is the methodology used in data oriented analysis.
• Architectures: descriptions of the items they define.
• IE methodology activities:
  1. Enterprise Analysis
  2. Business Area Analysis (BAA)
  3. Business System Design
  4. Construction
  5. Maintenance
• IE methodology:
  • ignores the current business organization, applications and procedures.
  • focuses on how the business should work, rather on how it currently does work.
  • reengineering is a common practice in IE methodology.

concepts and definitions

• relational database theory:
  • based on the mathematical set theory, or relational algebra.
  • programs do not do any read or write operations, instead DBMS does.
  • the result of operations is always a set.
• problems with DFDs:
  • DFD does not accommodate time
  • DFD has no sequence to processing
  • DFD assign media to data early in the analysis
• problems of DFDs are solved by using Process Data Flow Diagrams (PDFD).

Business Area Analysis (BAA)

• tasks performed during BAA:
  1. data modeling:
     • ERD + structure analysis
  2. data analysis
  3. functional decomposition (process modeling)
  4. process dependency analysis
  5. process data flow diagram (PDFD)
  6. process/data interaction mapping and analysis
  7. building data dictionary or repository for all information found during BAA
• entity:
  • object or concept that is important to the business.
  • three kinds:
    1. functional entity: independent of all other entities, can be defined without thinking about other entities. eg customer, product, order.
    2. attributive entity: entity that its existence depends on another entity or contain repetitive information relating to fundamental entity. eg customer address, order item, you would not have an order without items, but you can not have an order item without an order.
    3. associative entity: entity that connects two other entities. eg customer order, customer product. aka, join tables.
• Normalization is the refinement of data relationships to remove repeating information, partial key dependencies, and nonkey dependencies.
• Normalization methods:
  1. direct method:
     1. form 1:n relationships: create attributive entities.
     2. form m:n relationships: create associative entities or relationship entities (new intermediate join entity), aka, results in two 1:m relationships.
  2. tubular method:
     • recommended when data and relationships are not well defined.
     • define entities and relationships clearly; remove any dependencies that does not rely on the primary key
     • first form: remove repeating information, create attributive entities for 1:n relationships, associate entities for m:n relationships.
     • second form: remove partial key dependencies.
     • third form: remove nonkey dependencies.
     • the result is fundamental entities and relationship entities.
• example of normalized ERD https://imgur.com/YCNyPYv.png
• Entity structure analysis:
  • performed to determine weather a class structure applies, should we create class for this entity or not.
• business functional decomposition then business functions analysis are performed.
  • activity is some procedure in a business function. one activity has many processes
  • activity is a verb. eg create order, process order, ship order. process is a noun. eg Accounting, marketing.
  • function -> activity -> process -> elementary process.
  • elementary process is the smallest unit of work users identify
• functions decomposed and analyzed to create process dependency diagram:
  • process relationships are logical connections between processes than contain cardinality, sequence, iteration and selection components.
• process dependency diagram is expanded to process data flow diagram PDFD:
  • connections are due to passing data from one process to another as process data trigger.
  • triggers can be: data, events.
  • differences between DFD and PDFD:
    • PDFD has sequence to processing and time.
    • PDFD explicitly shows the media that connects processes.
• last step is to create an entity/process matrix, also known as a CRUD matrix.

Develop ERD

• relationships do NOT describe processes or processing. they describe ownership.
• 1:n relationships are:
  • customer -<- order, customer has n orders, order has 1 customer.
  • 1 side has many from the n side, n side has one and only one from the 1 side.
• any side can be required or optional.
• Required and optional in ERd:
• required and optional in ERD