Entity Relational Modeling
vs.
Dimensional Modeling

Strengths of Dimensional modeling

• Predictable
  
  
• Withstands unexpected Changes in user behavior
  
  
• Accommodate unexpected data element
  
  
• Handling Slowly Changing Dimensions
  
• Mange and use aggregates
  
• Snow Flaking dimensional Modeling

Usage

• E-R is good for reporting and point queries while dimensional modeling is good for ad-hoc query analysis.
  
  
• Many times, E-R translates to an E-R based data warehouse and a dimensional data mart layer.
  
  
• Dimensional imposes some rules on the modeling, but results in a model that has the access methods inherent by virtue of the relationships.
  
  
• Users are also better able to relate to the 'see measure by dimensional value(s)' paradigm than 'anything goes'.
  
  
• Dimensional is hard to come by directy from source, so in this approch dimensional is probably supported by some E-R, normalized tables.
  
  
• Also, there should be flexibility in the mart architecture such that marts are designed 'for purpose' - not all marts are for ad-hoc query analysis or generalized purposes and thus, not all marts should be dimensional.

Back to Entity Relational Modeling   vs.  Dimensional Modeling Index